Advertisement

Clinical and CT patterns to predict EGFR mutation in patients with non-small cell lung cancer: A systematic literature review and meta-analysis

Open AccessPublished:February 07, 2022DOI:https://doi.org/10.1016/j.ejro.2022.100400

      Highlights

      • GGO, air bronchogram, vascular convergence, pleural retraction, and spiculated margins, are risk factors for EGFR mutation.
      • Early disease stage, female gender and non-smoking status are risk factors for EGFR mutation.
      • Cavitation is a protective factor for EGFR mutation.

      Abstract

      Purpose

      This study aims to determine if the presence of specific clinical and computed tomography (CT) patterns are associated with epidermal growth factor receptor (EGFR) mutation in patients with non-small cell lung cancer.

      Methods

      A systematic literature review and meta-analysis was carried out in 6 databases between January 2002 and July 2021. The relationship between clinical and CT patterns to detect EGFR mutation was measured and pooled using odds ratios (OR). These results were used to build several mathematical models to predict EGFR mutation.

      Results

      34 retrospective diagnostic accuracy studies met the inclusion and exclusion criteria. The results showed that ground-glass opacities (GGO) have an OR of 1.86 (95%CI 1.34 −2.57), air bronchogram OR 1.60 (95%CI 1.38 – 1.85), vascular convergence OR 1.39 (95%CI 1.12 – 1.74), pleural retraction OR 1.99 (95%CI 1.72 – 2.31), spiculation OR 1.42 (95%CI 1.19 – 1.70), cavitation OR 0.70 (95%CI 0.57 – 0.86), early disease stage OR 1.58 (95%CI 1.14 – 2.18), non-smoker status OR 2.79 (95%CI 2.34 – 3.31), female gender OR 2.33 (95%CI 1.97 – 2.75). A mathematical model was built, including all clinical and CT patterns assessed, showing an area under the curve (AUC) of 0.81.

      Conclusions

      GGO, air bronchogram, vascular convergence, pleural retraction, spiculated margins, early disease stage, female gender, and non-smoking status are significant risk factors for EGFR mutation. At the same time, cavitation is a protective factor for EGFR mutation. The mathematical model built acts as a good predictor for EGFR mutation in patients with lung adenocarcinoma.

      Abbreviations:

      NSCLC (non-small cell lung carcinoma), EGFR (epidermal growth factor), KRAS (Kirsten rat sarcoma viral oncogene homolog), ALK (anaplastic lymphoma kinase mutation), EGFR TKI (epidermal growth factor receptor tyrosine kinase inhibitors), GGO (Ground glass opacities), OR (Odds ratios), CT (computed tomography), AUC (area under the curve), PRISMA (Preferred Reporting Items for Systematic Review and Meta-analysis), TP (True Positive), TN (True Negative), FP (False positives), FN (False negatives), ROC (Receiver Operating Characteristics), QUADAS-2 (Quality Assessment of Diagnostic Accuracy Studies-2)

      Keywords

      1. Introduction

      Lung cancer is a frequent non-hematological malignancy that represents 18% of all neoplasms, affecting mainly men with a median age of 70 years, and representing the leading cause of cancer-related mortality in 2020, with 1.8 million deaths [
      • de Groot P.M.
      • Wu C.C.
      • Carter B.W.
      • Munden R.F.
      The epidemiology of lung cancer.
      ,
      • Sung H.
      • Ferlay J.
      • Siegel R.L.
      • Laversanne M.
      • Soerjomataram I.
      • Jemal A.
      • Bray F.
      BF. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries.
      ]. Lung neoplasms are traditionally divided into non-small cell lung carcinoma (NSCLC) and small cell lung carcinoma [
      • Herbst R.S.
      • Heymach J.V.
      • Lippman S.M.
      Lung cancer.
      ]. Approximately 85% of lung neoplasms are NSCLC, comprising three main subtypes: adenocarcinoma, squamous cell carcinoma, and large cell carcinoma, which can be affected by mutations in epidermal growth factor receptor (EGFR), Kirsten rat sarcoma viral oncogene homolog (KRAS), and anaplastic lymphoma kinase mutation (ALK) [
      • Herbst R.S.
      • Heymach J.V.
      • Lippman S.M.
      Lung cancer.
      ,
      • Avrillon V.
      • Pérol M.
      Alectinib for treatment of ALK-positive non-small-cell lung cancer.
      ,
      • Bareschino M.A.
      • Schettino C.
      • Rossi A.
      • Maione P.
      • Sacco P.C.
      • Zeppa R.
      • et al.
      Treatment of advanced non small cell lung cancer.
      ]. The detection of EGFR mutation in patients with NSCLC has gained relevance in the last years due to the development of epidermal growth factor receptor tyrosine kinase inhibitors (EGFR TKI), which have shown better outcomes than chemotherapy in selected patients [
      • Lemjabbar-Alaoui H.
      • Hassan O.U.
      • Yang Y.-W.
      • Buchanan P.
      Lung cancer: biology and treatment options.
      ,
      • Shepherd F.A.
      • Rodrigues Pereira J.
      • Ciuleanu T.
      • Tan E.H.
      • Hirsh V.
      • Thongprasert S.
      • Campos D.
      • Maoleekoonpiroj S.
      • Smylie M.
      • Martins R.
      • van Kooten M.
      • Dediu M.
      • Findlay B.
      • Tu D.
      • Johnston D.
      • Bezjak A.
      • Clark G.
      • Santabárbara P.
      • Seymour L.
      Erlotinib in previously treated non-small-cell lung cancer.
      ,
      • Yoneda K.
      • Imanishi N.
      • Ichiki Y.
      • Tanaka F.
      Treatment of non-small cell lung cancer with EGFR-mutations.
      ,
      • Liu Y.
      • Kim J.
      • Qu F.
      • Liu S.
      • Wang H.
      • Balagurunathan Y.
      • Ye Z.
      • Gillies R.J.
      CT features associated with epidermal growth factor receptor mutation status in patients with lung adenocarcinoma.
      ].
      In some scenarios, it is not feasible to acquire adequate tissue for EGFR mutation analysis due to patients' inoperability, small tissue in the biopsy, or sampling artifacts [
      • Liu Y.
      • Kim J.
      • Qu F.
      • Liu S.
      • Wang H.
      • Balagurunathan Y.
      • Ye Z.
      • Gillies R.J.
      CT features associated with epidermal growth factor receptor mutation status in patients with lung adenocarcinoma.
      ]. To deal with this issue, certain predictor factors for EGFR mutation such, as non-smoking status, Asian ethnicity, and female gender, have been proposed; nonetheless, these are not enough to guide the treatment [
      • Ramlau R.
      • Krawczyk P.
      • Dziadziuszko R.
      • Chmielewska I.
      • Milanowski J.
      • Olszewski W.
      • Stencel K.
      • Ramlau-Piątek K.
      • Segiet A.
      • Skroński M.
      • Grudny J.
      • Chorostowska-Wynimko J.
      Predictors of EGFR mutation and factors associated with clinical tumor stage at diagnosis: experience of the INSIGHT study in Poland.
      ,
      • Aye P.S.
      • Tin Tin S.
      • McKeage M.J.
      • Khwaounjoo P.
      • Cavadino A.
      • Elwood J.M.
      Development and validation of a predictive model for estimating EGFR mutation probabilities in patients with non-squamous non-small cell lung cancer in New Zealand.
      ]. For that reason, the association between chest computed tomography (CT) patterns and EGFR mutation in NSCLC has been a topic of active research in the last years due to its potential to predict the therapeutic efficacy of EGFR TKI in patients whose tissue samples could not be obtained successfully. However, the evidence supporting CT patterns associated with EGFR mutation is contradictory, indicating the presence of a knowledge gap [
      • Lv J.
      • Zhang H.
      • Ma J.
      • Ma Y.
      • Gao G.
      • Song Z.
      • Yang Y.
      Comparison of CT radiogenomic and clinical characteristics between EGFR and KRAS mutations in lung adenocarcinomas.
      ,
      • Zhang H.
      • Cai W.
      • Wang Y.
      • Liao M.
      • Tian S.
      CT and clinical characteristics that predict risk of EGFR mutation in non-small cell lung cancer: a systematic review and meta-analysis.
      ,
      • Cheng Z.
      • Shan F.
      • Yang Y.
      • Shi Y.
      • Zhang Z.
      CT characteristics of non-small cell lung cancer with epidermal growth factor receptor mutation: a systematic review and meta-analysis.
      ]. Therefore, this systematic literature review and meta-analysis aims to determine if clinical and chest CT patterns such as ground-glass opacities (GGO), air bronchogram, vascular convergence, pleural retraction, spiculated margins, cavitation, early disease stage, female gender, and non-smoker status are risk factors for EGFR mutation in patients with NSCLC.

      2. Material and methods

      This systematic literature review and meta-analysis was performed based on the Preferred Reporting Items for Systematic Review and Meta-analysis (PRISMA) methodology.

      2.1 Eligibility criteria

      We searched for articles published in journals between January 2002 and July 2021. The lower margin of the search dates was set this way because 2002 was when EGFR mutation was first discovered in a subset of lung cancers [
      • Mitsudomi T.
      • Yatabe Y.
      Epidermal growth factor receptor in relation to tumor development: EGFR gene and cancer.
      ]; therefore, it is not expected to have articles seeking EFGR mutation in lung cancer before that date.
      The inclusion criteria were cross-sectional studies, comparative studies, retrospective studies, randomized and non-randomized clinical trials, articles published in English, Spanish or French, performed in patients all around the world, either hospitalized or ambulatory with a diagnosis of NSCLC presenting a disease progression at any stage, no age or gender predilection were set. The intervention performed must be CT, in which the image patterns had to be characterized and then compared against biopsy or cytology to detect EGFR mutation.
      We excluded articles that did not provide information regarding the True Positive (TP), True Negative (TN), False Positives (FP), and False Negatives (FN) of the different clinical and CT patterns to diagnose EGFR mutation, papers with a high risk of bias based on the Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) tool and articles that assessed patients who received radiotherapy, chemotherapy, biological therapy, or surgery before CT.

      2.2 Sources of information

      The literature search was performed in several databases, including the Medical Literature Analysis and Retrieval System Online (MEDLINE), the Excerpta Medica database (EMBASE), SCOPUS, The Virtual Health Library (VHL), and the African Index Medicus. These databases were chosen because they summarize the most important publications in different regions worldwide. Also, a search in google scholar was carried out to identify articles not published in major databases. Secondary searching methods were applied, such as the snowballing technique, in which the references of the articles included in the meta-analysis were reviewed to identify more papers. Additionally, we decided to perform a hand search in three major radiology journals (Radiology, European Journal of Radiology, American Journal of Roentgenology); we have agreed to hand search these journals because they were not shown substantially in the database search, and they have a high impact factor in radiology.

      2.3 Search

      Mesh terms, keywords, and synonyms were used to guarantee that no articles were missed. The search strategy applied was: (((Epidermal growth factor receptor) OR EGFR) OR epidermal growth factor receptor mutation) AND (((((Carcinoma, Non-Small-Cell Lung) OR NSCLC) OR non-small cell lung cancer) OR lung adenocarcinoma) OR lung cancer) OR lung carcinoma) AND (((Tomography, X-Ray Computed) OR Tomography, Spiral Computed) OR computed tomography) OR CT) AND ((Biopsy) OR cytology).

      2.4 Study selection

      All the articles found through the search strategy were blindly uploaded to Mendeley; subsequently, the duplicates were removed using the Mendeley duplicate detection tool [
      • Herrera Ortiz A.F.
      • Fernández Beaujon L.J.
      • García Villamizar S.Y.
      • Fonseca López F.F.
      Magnetic resonance versus computed tomography for the detection of retroperitoneal lymph node metastasis due to testicular cancer: a systematic literature review.
      ]. When the duplicates were removed, the screening process was performed based on title and abstract. The pre-selected articles were assessed for inclusion and exclusion criteria using prespecified questions; then, the articles that met the selection criteria were read completely and subjected to a quality assessment. Two authors carried out the entire process independently, and if disagreements were presented, these were resolved by consensus. If two articles used the same cohort of patients, the authors would only include the study with the largest sample size to avoid enrolling the same individuals twice.

      2.5 Data extraction and missing data

      All five authors extracted the information from each article, registering it in a qualitative data extraction table which contained the author's names, year of publication, country of publication, study type, number of patients included, the mean age of the participants, the NSCLC histopathological subtype, the disease stage of the participants, the index test (CT), the reference standard used (biopsy or cytology), and the CT patterns described.
      Additionally, a quantitative data extraction table was developed to register the TP, FP, FN, and TN of the different clinical and CT patterns when compared to biopsy or cytology to detect EGFR mutation.
      When important information was missing in any article selected to be included, the paper's corresponding author was contacted to provide the missing data. Finally, the information was left empty if the corresponding author could not be reached.
      Two authors reviewed all the information extracted in the qualitative and quantitative data extraction table to guarantee that no typos or inaccurate data were included. In case of disagreements, these were resolved by consensus.

      2.6 Outcomes

      The primary outcome was to evaluate the association of specific clinical characteristics (Non-smoker status, female gender, early disease stage) and chest CT patterns (GGO, air bronchogram, pleural retraction, vascular convergence, spiculation, and cavitation) with EGFR mutation determined by biopsy or cytology in patients with NSCLC.
      The secondary outcome was to build several mathematical models, including these CT and clinical patterns, to predict the presence of EGFR mutation in patients with NSCLC.

      2.7 Quality assessment

      Due to the nature of our research question, all the articles included in the systematic literature review and meta-analysis were of diagnostic accuracy; therefore, the QUADAS-2 tool for diagnostic studies was used (https://www.bristol.ac.uk/population-health-sciences/projects/quadas/quadas-2/). Only the studies that presented a low or acceptable risk of bias with little concern regarding applicability were considered for inclusion.
      The QUADAS-2 tool is a checklist that has four domains, distributed in the following way: 1. "Patient selection," 2. "Index test," 3. "Reference Standard," 4. "Flow and timing." Each domain is evaluated regarding bias, and the first three domains are additionally assessed regarding their applicability. Two independent authors performed the quality assessment of each article, and if disagreements were presented, these were resolved by consensus. The "traffic light" plot for QUADAS-2 was created using the Robvis tool [
      • McGuinness L.A.
      • Higgins J.P.T.
      Risk-of-bias VISualization (robvis): an R package and Shiny web app for visualizing risk-of-bias assessments.
      ].

      2.8 Effect measures

      Because the primary outcome was categorical, each article's TP, FP, FN, and TN values were used to calculate odds ratios (OR), which would then be pooled in a quantitative synthesis. The secondary outcomes were expressed in relative frequencies.

      2.9 Statistical analysis

      All statistical analyses were performed using STATA 17 (StataCorp LLC, College Station, TX).
      The statistical heterogeneity of the studies included was explored for each clinical and CT pattern using Cochran's Q test and I2 test. If the Cochran's Q test was < 0.05, we considered that the meta-analysis presented a high degree of heterogeneity. If the I2 test value was < 50%, we used a fixed-effect model; nevertheless, we used a random-effect model if the I2 test value was > 50% [
      • Ortiz A.F.H.
      • Camacho E.C.
      • Rojas J.C.
      • Camacho T.C.
      • Guevara S.Z.
      • Cuenca N.T.R.
      A practical guide to perform a systematic literature review and meta-analysis.
      ]. If discrepancies between the Cochran’s Q test and the I2 tests were presented, a random effect model was chosen.
      A forest plot based on OR was performed for each clinical and CT pattern to pool the effects of all the articles. Publication bias was assessed with the egger's test for each forest plot; we considered positive for publication bias if the p-value was < 0.05.
      A sensitivity analysis was carried out for each forest plot to evaluate the robustness of the results.
      The TP, FP, FN, TN of the CT patterns and clinical characteristics with statistically significant results were used to build several mathematical models using logistic regression. Based on these models, we created multiple Receiver Operating Characteristics (ROC) curves, from which the Area Under the Curve (AUC) was calculated.

      3. Results

      3.1 Search results

      The search provided a total of 1202 non-duplicated citations screened based on title and abstract, from which 1152 did not match the research question. Leaving a total of 50 articles that were read completely, identifying that 4 did not compare EGFR status against CT, 10 provided insufficient details to calculate the OR, 1 used the same population of an already included article but with another research question, and 1 had limited rigor based on QUADAS-2 tool, leaving a total of 34 papers that were included in the systematic literature review and meta-analysis. These results are better schematized in Fig. 1.
      Fig. 1
      Fig. 1PRISMA flow diagram, CT: Computed tomography.

      3.2 Summary of studies

      A total of 34 retrospective diagnostic accuracy studies were included in the final analysis [
      • Liu Y.
      • Kim J.
      • Qu F.
      • Liu S.
      • Wang H.
      • Balagurunathan Y.
      • Ye Z.
      • Gillies R.J.
      CT features associated with epidermal growth factor receptor mutation status in patients with lung adenocarcinoma.
      ,
      • Lv J.
      • Zhang H.
      • Ma J.
      • Ma Y.
      • Gao G.
      • Song Z.
      • Yang Y.
      Comparison of CT radiogenomic and clinical characteristics between EGFR and KRAS mutations in lung adenocarcinomas.
      ,
      • Hsu J.S.
      • Huang M.S.
      • Chen C.Y.
      • Liu G.C.
      • Liu T.C.
      • Chong I.W.
      • Chou S.H.
      • Yang C.J.
      Correlation between EGFR mutation status and computed tomography features in patients with advanced pulmonary adenocarcinoma.
      ,
      • Zhang G.
      • Cao Y.
      • Zhang J.
      • Ren J.
      • Zhao Z.
      • Zhang X.
      • Li S.
      • Deng L.
      • Zhou J.
      Predicting EGFR mutation status in lung adenocarcinoma: development and validation of a computed tomography-based radiomics signature.
      ,
      • Qin X.
      • Gu X.
      • Lu Y.
      • Zhou W.
      EGFR-TKI-sensitive mutations in lung carcinomas: are they related to clinical features and CT findings?.
      ,
      • Rizzo S.
      • Petrella F.
      • Buscarino V.
      • De Maria F.
      • Raimondi S.
      • Barberis M.
      • Fumagalli C.
      • Spitaleri G.
      • Rampinelli C.
      • De Marinis F.
      • Spaggiari L.
      • Bellomi M.
      CT radiogenomic characterization of EGFR, K-RAS, and ALK mutations in non-small cell lung cancer.
      ,
      • Hasegawa M.
      • Sakai F.
      • Ishikawa R.
      • Kimura F.
      • Ishida H.
      • Kobayashi K.C.T.
      Features of epidermal growth factor receptor-mutated adenocarcinoma of the lung: comparison with nonmutated adenocarcinoma.
      ,
      • Mori M.
      • Hayashi H.
      • Fukuda M.
      • Honda S.
      • Kitazaki T.
      • Shigematsu K.
      • Matsuyama N.
      • Otsubo M.
      • Nagayasu T.
      • Hashisako M.
      • Tabata K.
      • Uetani M.
      • Ashizawa K.
      Clinical and computed tomography characteristics of non-small cell lung cancer with ALK gene rearrangement: comparison with EGFR mutation and ALK/EGFR-negative lung cancer.
      ,
      • Glynn C.
      • Zakowski M.F.
      • Ginsberg M.S.
      Are there imaging characteristics associated with epidermal growth factor receptor and KRAS mutations in patients with adenocarcinoma of the lung with bronchioloalveolar features?.
      ,
      • Dai J.
      • Shi J.
      • Soodeen-Lalloo A.K.
      • Zhang P.
      • Yang Y.
      • Wu C.
      • Jiang S.
      • Jia X.
      • Fei K.
      • Jiang G.
      Air bronchogram: a potential indicator of epidermal growth factor receptor mutation in pulmonary subsolid nodules.
      ,
      • Zhao J.
      • Dinkel J.
      • Warth A.
      • Penzel R.
      • Reinmuth N.
      • Schnabel P.
      • Muley T.
      • Meister M.
      • Zabeck H.
      • Steins M.
      • Yang J.
      • Zhou Q.
      • Schlemmer H.P.
      • Herth F.J.F.
      • Kauczor H.U.
      • Heussel C.P.
      CT characteristics in pulmonary adenocarcinoma with epidermal growth factor receptor mutation.
      ,
      • Yang X.
      • Dong X.
      • Wang J.
      • Li W.
      • Gu Z.
      • Gao D.
      • Zhong N.
      • Guan Y.
      Computed tomography‐based radiomics signature: a potential indicator of epidermal growth factor receptor mutation in pulmonary adenocarcinoma appearing as a subsolid nodule.
      ,
      • Choi C.M.C.-M.C.-M.
      • Kim M.Y.M.Y.
      • Hwang H.J.H.J.H.J.
      • Lee J.B.J.B.
      • Kim W.S.W.S.
      Advanced adenocarcinoma of the lung: comparison of CT characteristics of patients with anaplastic lymphoma kinase gene rearrangement and those with epidermal growth factor receptor mutation.
      ,
      • Yano M.
      • Sasaki H.
      • Kobayashi Y.
      • Yukiue H.
      • Haneda H.
      • Suzuki E.
      • Endo K.
      • Kawano O.
      • Hara M.
      • Fujii Y.
      Epidermal growth factor receptor gene mutation and computed tomographic findings in peripheral pulmonary adenocarcinoma.
      ,
      • Suh Y.J.
      • Lee H.J.
      • Kim Y.J.
      • Kim K.G.
      • Kim H.
      • Jeon Y.K.
      • Kim Y.T.
      Computed tomography characteristics of lung adenocarcinomas with epidermal growth factor receptor mutation: a propensity score matching study.
      ,
      • Tu W.
      • Sun G.
      • Fan L.
      • Wang Y.
      • Xia Y.
      • Guan Y.
      • Li Q.
      • Zhang D.
      • Liu S.
      • Li Z.
      Radiomics signature: a potential and incremental predictor for EGFR mutation status in NSCLC patients, comparison with CT morphology.
      ,
      • Hong S.J.
      • Kim T.J.
      • Choi Y.W.
      • Park J.S.
      • Chung J.H.L.K.
      Radiogenomic correlation in lung adenocarcinoma with epidermal growth factor receptor mutations: Imaging features.
      ,
      • Lee Y.
      • Lee H.J.
      • Kim Y.T.
      • Kang C.H.
      • Goo J.M.
      • Park C.M.
      • Paeng J.C.
      • Chung D.H.
      • Jeon Y.K.
      Imaging characteristics of stage I non-small cell lung cancer on CT and FDG-PET: relationship with epidermal growth factor receptor protein expression status and survival.
      ,
      • Yang Y.
      • Yang Y.
      • Zhou X.
      • Song X.
      • Liu M.
      • He W.
      • Wang H.
      • Wu C.
      • Fei K.
      • Jiang G.
      EGFR L858R mutation is associated with lung adenocarcinoma patients with dominant ground-glass opacity.
      ,
      • Zou J.
      • Lv T.
      • Zhu S.
      • Lu Z.
      • Shen Q.
      • Xia L.
      • Wu J.
      • Song Y.
      • Liu H.
      Computed tomography and clinical features associated with epidermal growth factor receptor mutation status in stage I/II lung adenocarcinoma.
      ,
      • Hua Wang Matthew B.
      • Schabath Ying
      • Liu Ying
      • Han Qi
      • Li
      • Robert J.
      Gillies et al. Clinical and CT characteristics of surgically resected lung adenocarcinomas harboring ALK rearrangements or EGFR mutations.
      ,
      • Kim T.J.
      • Lee C.T.
      • Jheon S.H.
      • Park J.S.
      • Chung J.H.
      Radiologic characteristics of surgically resected non-small cell lung cancer with ALK rearrangement or EGFR mutations.
      ,
      • Park J.
      • Kobayashi Y.
      • Urayama K.Y.
      • Yamaura H.
      • Yatabe Y.
      • Hida T.
      Imaging characteristics of driver mutations in EGFR, KRAS, and ALK among treatment-naïve patients with advanced lung adenocarcinoma.
      ,
      • Zhou J.Y.
      • Zheng J.
      • Yu Z.F.
      • Xiao W.B.
      • Zhao J.
      • Sun K.
      • Wang B.
      • Chen X.
      • Jiang L.N.
      • Ding W.
      • Zhou J.Y.
      Comparative analysis of clinicoradiologic characteristics of lung adenocarcinomas with ALK rearrangements or EGFR mutations.
      ,
      • Shi Z.
      • Zheng X.
      • Shi R.
      • Song C.
      • Yang R.
      • Zhang Q.
      • Wang X.
      • Lu J.
      • Yu Y.
      • Jiang T.
      Score for lung adenocarcinoma in China with EGFR mutation of exon 19 Combination of clinical and radiological characteristics analysis.
      ,
      • Woo J.H.
      • Kim T.J.
      • Kim T.S.
      • Han J.
      CT features and disease spread patterns in ROS1-rearranged lung adenocarcinomas: comparison with those of EGFR-mutant or ALK-rearranged lung adenocarcinomas.
      ,
      • Hsu K.H.
      • Chen K.C.
      • Yang T.Y.
      • Yeh Y.C.
      • Chou T.Y.
      • Chen H.Y.
      • Tsai C.R.
      • Chen C.Y.
      • Hsu C.P.
      • Hsia J.Y.
      • Chuang C.Y.
      • Tsai Y.H.
      • Chen K.Y.
      • Huang M.S.
      • Su W.C.
      • Chen Y.M.
      • Hsiung C.A.
      • Chang G.C.
      • Chen C.J.
      • Yang P.C.
      Epidermal growth factor receptor mutation status in stage I lung adenocarcinoma with different image patterns.
      ,
      • Shi Z.
      • Zheng X.
      • Shi R.
      • Song C.
      • Yang R.
      • Zhang Q.
      • Wang X.
      • Lu J.
      • Yu Y.
      • Liu Q.
      • Jiang T.
      Radiological and clinical features associated with epidermal growth factor receptor mutation status of exon 19 and 21 in lung adenocarcinoma.
      ,
      • Cao Y.
      • Xu H.
      A new predictive scoring system based on clinical data and computed tomography features for diagnosing EGFR-mutated lung adenocarcinoma.
      ,
      • Lee H.J.
      • Kim Y.T.
      • Kang C.H.
      • Zhao B.
      • Tan Y.
      • Schwartz L.H.
      • Persigehl T.
      • Jeon Y.K.
      • Chung D.H.
      Epidermal growth factor receptor mutation in lung adenocarcinomas: relationship with CT characteristics and histologic subtypes.
      ,
      • Sugano M.
      • Shimizu K.
      • Nakano T.
      • Kakegawa S.
      • Miyamae Y.
      • Kaira K.
      • Araki T.
      • Kamiyoshihara M.
      • Kawashima O.
      • Takeyoshi I.
      Correlation between computed tomography findings and epidermal growth factor receptor and KRAS gene mutations in patients with pulmonary adenocarcinoma.
      ,
      • Sacconi B.
      • Anzidei M.
      • Leonardi A.
      • Boni F.
      • Saba L.
      • Scipione R.
      • Anile M.
      • Rengo M.
      • Longo F.
      • Bezzi M.
      • Venuta F.
      • Napoli A.
      • Laghi A.
      • Catalano C.
      Analysis of CT features and quantitative texture analysis in patients with lung adenocarcinoma: a correlation with EGFR mutations and survival rates.
      ,
      • Chen Y.
      • Yang Y.
      • Ma L.
      • Zhu H.
      • Feng T.
      • Jiang S.
      • Wei Y.
      • Wang T.
      • Sun X.
      Prediction of EGFR mutations by conventional CT-features in advanced pulmonary adenocarcinoma.
      ,
      • Sabri A.
      • Batool M.
      • Xu Z.
      • Bethune D.
      • Abdolell M.
      • Manos D.
      Predicting EGFR mutation status in lung cancer: proposal for a scoring model using imaging and demographic characteristics.
      ]. These articles assessed 10355 individuals; the sample size fluctuated between 64 and 864 patients per study. The mean age of the individuals ranged between 56 and 68 years. The qualitative synthesis of the articles included is presented in Table 1. If readers would like to access the crude data, this is attached in Appendix A.
      Table 1Qualitative synthesis of the articles included.
      Gender
      Male4501/10355 patients
      Female4464/10355 patients
      Not described1390/10355 patients
      EGFR status
      EGFR positive5046/10355 patients
      EGFR negative5309/10355 patients
      Smoking status
      Smoker3244/10355 patients
      Never smoked4970/10355 patients
      Not described2141/10355 patients
      Country
      China6254/10355 patients
      Korea2046/10355 patients
      Japan926/10355 patients
      Italy353/10355 patients
      Taiwan311/10355 patients
      Germany282/10355 patients
      Canada119/10355 patients
      United States64/10355 patients
      Disease stage
      Stage I2507/10355 patients
      Stage II562/10355 patients
      Stage III849/10355 patients
      Stage IV1396/10355 patients
      Not described5041/10355 patients
      Histological subtype
      Adenocarcinoma10079/10355 patients
      Squamous-cell carcinoma139/10355 patients
      Large-cell carcinoma2/10355 patients
      Not clearly described135/10355 patients
      CT pattern evaluated
      GGO6893/10355 patients
      Air bronchogram7630/10355 patients
      Vascular convergence1716/10355 patients
      Pleural retraction3471/10355 patients
      Spiculation5871/10355 patients
      Cavitation4891/10355 patients
      Sample acquisition method
      Biopsy10073/10355 patients
      Cytology282/10355 patients
      EGFR mutation analysis test
      PCR8922/10355 patients
      FISH198/10355 patients
      Immunohistochemistry214/10355 patients
      Other850/10355 patients
      Not described171/10355 patients
      Interpretation of the images
      Radiologists or clinicians with experience8246/10355 patients
      Machine learning tools2109/10355 patients

      3.3 Risk of bias within studies

      All the studies included in the analysis showed a low or acceptable risk of bias according to the QUADAS-2 tool. The findings are presented in Fig. 2. If readers want to access the quality assessment of each article, this is attached in Appendix B.
      Fig. 2
      Fig. 2Quality assessment of all the articles included in the meta-analysis.

      3.4 GGO and EGFR mutation

      A total of 6893 patients from 23 different studies were pooled to evaluate the association between GGO and EGFR mutation. The Cochrane Q test p-value was = 0.000, and the I2 value was = 80.3%; based on these results, we considered high heterogeneity in the data; therefore, a random effect model was performed. The overall effect showed an OR of 1.86 (95% CI 1.34 −2.57) (Fig. 3).
      Fig. 3
      Fig. 3Forest plot for GGO and EGFR mutation.

      3.5 Air bronchogram and EGFR mutation

      7630 patients from 27 different studies were pooled to evaluate the association between air bronchogram and EGFR mutation. The Cochrane Q test p-value was = 0.007, and the I2 value was = 44.7%; therefore, we considered a moderate heterogeneity in the data; for which a random effect model was performed. The overall effect showed an OR of 1.60 (95% CI 1.38 – 1.85) (Fig. 4).
      Fig. 4
      Fig. 4Forest plot for air bronchogram and EGFR mutation.

      3.6 Vascular convergence and EGFR mutation

      1716 patients from 6 different studies were pooled to evaluate the association between vascular convergence and EGFR mutation. The Cochrane Q test p-value was = 0.845, and the I2 value was = 0%. Based on these results, we considered low heterogeneity in the data; therefore, a fixed-effect model was carried out. The overall effect showed an OR of 1.39 (95% CI 1.12 – 1.74) (Fig. 5).
      Fig. 5
      Fig. 5Forest plot for vascular convergence and EGFR mutation.

      3.7 Pleural retraction and EGFR mutation

      3471 patients from 11 different studies were pooled to evaluate the association between pleural retraction and EGFR mutation. The Cochrane Q test p-value was= 0.498, and the I2 value was = 0%; therefore, we considered low heterogeneity in the data, for which a fixed-effect model was carried out. The overall effect showed an OR of 1.99 (95% CI 1.72 – 2.31) (Fig. 6).
      Fig. 6
      Fig. 6Forest plot for pleural retraction and EGFR mutation.

      3.8 Spiculation and EGFR mutation

      A total of 5871 patients from 21 different studies were pooled to evaluate the association between spiculated margins and EGFR mutation. The Cochrane Q test p-value was= 0.004, and the I2 value was = 51.2%. Based on these results, we considered moderate-high heterogeneity in the data; therefore, a random effect model was carried out. The overall effect showed an OR of 1.42 (95% CI 1.19 – 1.70) (Fig. 7).
      Fig. 7
      Fig. 7Forest plot for spiculated margins and EGFR mutation.

      3.9 Cavitation and EGFR mutation

      4891 patients from 15 different studies were pooled to evaluate the association between tumor cavitation and EGFR mutation. The Cochrane Q test p-value was= 0.759, and the I2 value was = 0%. Based on these results, we considered low heterogeneity in the data; therefore, a fixed-effect model was carried out. The overall effect showed an OR of 0.70 (95% CI 0.57 – 0.86) (Fig. 8).
      Fig. 8
      Fig. 8Forest plot for tumor cavitation and EGFR mutation.

      3.10 Early disease stage and EGFR mutation

      The disease stage was classified categorically as early-stage (I and II) and late-stage (III and IV). We pooled 2494 patients from 10 different studies to evaluate the association between early disease stage and EGFR mutation. The Cochrane Q test p-value was = 0.017 and the I2 value was = 55.2%. Based on these results, we considered moderate-high heterogeneity in the data; therefore, a random effect model was carried out. The overall effect showed an OR of 1.58 (95% CI 1.14 – 2.18) (Fig. 9).
      Fig. 9
      Fig. 9Forest plot for early disease stage and EGFR mutation.

      3.11 Non-smoker status and EGFR mutation

      8214 patients from 27 different studies were pooled to evaluate the association between non-smoker status and EGFR mutation. The Cochrane Q test p-value was = 0.000 and the I2 value was = 64.6%. Based on these results, we considered moderate-high heterogeneity in the data; therefore, a random effect model was carried out. The overall effect showed an OR of 2.79 (95% CI 2.34 – 3.31) (Fig. 10).
      Fig. 10
      Fig. 10Forest plot for non-smoker status and EGFR mutation.

      3.12 Female gender and EGFR mutation

      8965 patients from 31 different studies were pooled to evaluate the association between female gender and EGFR mutation. The Cochrane Q test p-value was= 0.000 and the I2 value was = 67.7%. Based on these results, we considered moderate-high heterogeneity in the data; therefore, a random effect model was carried out. The overall effect showed an OR of 2.33 (95% CI 1.97 – 2.75) (Fig. 11).
      Fig. 11
      Fig. 11Forest plot for female gender EGFR mutation.

      3.13 Sensitivity analysis

      Sensitivity analysis was performed for all the forest plots, showing that the overall effect did not cross over the null effect value at any moment, which indicates robustness of all the results. All the sensitivity analysis plots are shown in Appendix C.

      3.14 Publication bias

      Publication bias was assessed for all forest plots using the egger's test, which showed statistically significant results for the association between air bronchogram and EGFR mutation. Therefore, we considered publication bias, for which a funnel plot was performed, suggesting a lack of small studies with negative effects. The results of the egger's test are better depicted in Table 2.
      Table 2Results of the egger's test to assess publication bias.
      Outcome assessedP-value
      GGO and EGFR mutation0.17
      Air bronchogram and EGFR mutation0.0006
      Vascular convergence and EGFR mutation0.35
      Pleural retraction and EGFR mutation0.66
      Spiculation and EGFR mutation0.50
      Cavitation and EGFR mutation0.40
      Early disease stage and EGFR mutation0.77
      Non-smoker status and EGFR mutation0.80
      Female gender and EGFR mutation0.69

      3.15 Secondary outcomes

      All the clinical and CT characteristics assessed in this meta-analysis were used to develop 9 mathematical models to predict EGFR mutation. We performed a model including exclusively clinical factors such as female gender + non-smoker status + early disease stage, which showed an AUC of 0.63; however, when all the clinical and CT patterns assessed were added to the model, the AUC rises to 0.81 (Table 3). The ROC curve prediction for EGFR mutation based on all clinical and CT patterns assessed, is shown in Fig. 12.
      Table 3Mathematical model to predict EGFR mutation based on radiological and clinical data.
      Mathematical model to predict EGFR mutation based on radiological and clinical dataAUC
      Female gender + Non-smoker status + GGO + Air bronchogram + Vascular convergence + Cavitation+ Pleural retraction + Spiculation + Early disease stage0.81
      Female gender + Non-smoker status + Spiculation + Pleural retraction0.78
      Female gender + Non-smoker status + Spiculation0.71
      Female gender + Non-smoker status + Vascular convergence0.67
      Female gender + Non-smoker status + Pleural retraction0.67
      Female gender + Non-smoker status + Early disease stage0.63
      Female gender + Non-smoker status + Air bronchogram0.61
      Female gender + Non-smoker status + GGO0.60
      Female gender + Non-smoker status0.60
      Fig. 12
      Fig. 12ROC curve prediction for EGFR mutation based on female gender, non-smoker status, GGO, air bronchogram, vascular convergence, cavitation, pleural retraction, spiculation and early disease stage.

      4. Discussion

      This study aimed to identify the clinical and CT characteristics associated with EGFR mutation in patients with NSCLC, to develop a predictive model. We found that GGO, air bronchogram, vascular convergence, pleural retraction, non-smoker status, and female gender were significant risk factors for EGFR mutation in patients with lung adenocarcinoma, which is supported by previous literature [
      • Zhang H.
      • Cai W.
      • Wang Y.
      • Liao M.
      • Tian S.
      CT and clinical characteristics that predict risk of EGFR mutation in non-small cell lung cancer: a systematic review and meta-analysis.
      ]. However, this study showed results that have not been reported previously in other meta-analyses, such as the protective effects of cavitation for EGFR mutation and the association between spiculated margins and early disease stage with EGFR mutation. These discrepancies may be due to a lack of statistical power in previously published studies. This meta-analysis also showed that air bronchogram, spiculated margins, and GGO represents the most frequent CT patterns associated with lung adenocarcinoma.
      Mathematical models expressed as ROC curves involving CT and clinical characteristics to detect EGFR mutation have been published previously. Several authors have described models that showed an AUC ranging from 0.87 to 0.68 [
      • Liu Y.
      • Kim J.
      • Qu F.
      • Liu S.
      • Wang H.
      • Balagurunathan Y.
      • Ye Z.
      • Gillies R.J.
      CT features associated with epidermal growth factor receptor mutation status in patients with lung adenocarcinoma.
      ,
      • Zhang H.
      • Cai W.
      • Wang Y.
      • Liao M.
      • Tian S.
      CT and clinical characteristics that predict risk of EGFR mutation in non-small cell lung cancer: a systematic review and meta-analysis.
      ,
      • Zhang G.
      • Cao Y.
      • Zhang J.
      • Ren J.
      • Zhao Z.
      • Zhang X.
      • Li S.
      • Deng L.
      • Zhou J.
      Predicting EGFR mutation status in lung adenocarcinoma: development and validation of a computed tomography-based radiomics signature.
      ,
      • Glynn C.
      • Zakowski M.F.
      • Ginsberg M.S.
      Are there imaging characteristics associated with epidermal growth factor receptor and KRAS mutations in patients with adenocarcinoma of the lung with bronchioloalveolar features?.
      ,
      • Choi C.M.C.-M.C.-M.
      • Kim M.Y.M.Y.
      • Hwang H.J.H.J.H.J.
      • Lee J.B.J.B.
      • Kim W.S.W.S.
      Advanced adenocarcinoma of the lung: comparison of CT characteristics of patients with anaplastic lymphoma kinase gene rearrangement and those with epidermal growth factor receptor mutation.
      ,
      • Lee Y.
      • Lee H.J.
      • Kim Y.T.
      • Kang C.H.
      • Goo J.M.
      • Park C.M.
      • Paeng J.C.
      • Chung D.H.
      • Jeon Y.K.
      Imaging characteristics of stage I non-small cell lung cancer on CT and FDG-PET: relationship with epidermal growth factor receptor protein expression status and survival.
      ,
      • Woo J.H.
      • Kim T.J.
      • Kim T.S.
      • Han J.
      CT features and disease spread patterns in ROS1-rearranged lung adenocarcinomas: comparison with those of EGFR-mutant or ALK-rearranged lung adenocarcinomas.
      ]. Our meta-analysis found an AUC of 0.81 when combining all the clinical and CT patterns assessed, making this model a good predictor for EGFR mutation and subsequently a determinant for EGFR-TKI response. Some studies have shown that radiomic features can be useful to predict EGFR mutation; however, radiomic data combined with morphological characteristics have demonstrated to improve the predictive value just a little; therefore, this area continues to be an active research topic [
      • Liu Y.
      • Kim J.
      • Qu F.
      • Liu S.
      • Wang H.
      • Balagurunathan Y.
      • Ye Z.
      • Gillies R.J.
      CT features associated with epidermal growth factor receptor mutation status in patients with lung adenocarcinoma.
      ,
      • Liu Y.
      • Kim J.
      • Balagurunathan Y.
      • Li Q.
      • Garcia A.L.
      • Stringfield O.
      • Ye Z.
      • Gillies R.J.
      Radiomic features are associated with EGFR mutation status in lung adenocarcinomas.
      ].
      The likelihood of having EGFR mutation also varies depending on the histopathological subtype of the NSCLC tumor. For example, Song et al. described that adenocarcinomas with micropapillary or lepidic predominance were more prone to EGFR mutation [
      • Song Z.
      • Zhu H.
      • Guo Z.
      • Wu W.
      • Sun W.
      • Zhang Y.
      Correlation of EGFR mutation and predominant histologic subtype according to the new lung adenocarcinoma classification in Chinese patients.
      ]. Zhang et al. published that the presence of EGFR mutation was correlated with acinar predominant adenocarcinomas [
      • Zhang Y.
      • Sun Y.
      • Pan Y.
      • Li C.
      • Shen L.
      • Li Y.
      • Luo X.
      • Ye T.
      • Wang R.
      • Hu H.
      • Li H.
      • Wang L.
      • Pao W.
      • Chen H.
      Frequency of driver mutations in lung adenocarcinoma from female never-smokers varies with histologic subtypes and age at diagnosis.
      ]. On the other hand, Sun et al. showed that papillary predominant adenocarcinomas were more frequently associated with EGFR mutation [
      • Sekine A.
      • Tamura K.
      • Satoh H.
      • Tanaka T.
      • Tsunoda Y.
      • Tanaka T.
      • Takoi H.
      • Lin S.Y.
      • Yatagai Y.
      • Hashizume T.
      • Hayasihara K.
      • Saito T.
      Prevalence of underlying lung disease in smokers with epidermal growth factor receptor-mutant lung cancer.
      ].
      All EGFR mutation subtypes do not show the same radiological and clinical characteristics in patients with adenocarcinoma; for example, it has been described that EGFR mutation in exon 21 is shown more in non-smoker women and tumors with a higher proportion of GGO [
      • Hong S.J.
      • Kim T.J.
      • Choi Y.W.
      • Park J.S.
      • Chung J.H.L.K.
      Radiogenomic correlation in lung adenocarcinoma with epidermal growth factor receptor mutations: Imaging features.
      ,
      • Shi Z.
      • Zheng X.
      • Shi R.
      • Song C.
      • Yang R.
      • Zhang Q.
      • Wang X.
      • Lu J.
      • Yu Y.
      • Liu Q.
      • Jiang T.
      Radiological and clinical features associated with epidermal growth factor receptor mutation status of exon 19 and 21 in lung adenocarcinoma.
      ,
      • Lee H.J.
      • Kim Y.T.
      • Kang C.H.
      • Zhao B.
      • Tan Y.
      • Schwartz L.H.
      • Persigehl T.
      • Jeon Y.K.
      • Chung D.H.
      Epidermal growth factor receptor mutation in lung adenocarcinomas: relationship with CT characteristics and histologic subtypes.
      ]. Specifically, L858R mutation in exon 21 is frequently associated with broncho-alveolar adenocarcinoma and non-smoker status [
      • Yano M.
      • Sasaki H.
      • Kobayashi Y.
      • Yukiue H.
      • Haneda H.
      • Suzuki E.
      • Endo K.
      • Kawano O.
      • Hara M.
      • Fujii Y.
      Epidermal growth factor receptor gene mutation and computed tomographic findings in peripheral pulmonary adenocarcinoma.
      ,
      • Moriguchi H.
      • Kim T.Y.
      • Sato C.
      Gefitinib for refractory advanced non-small-cell lung cancer.
      ]. On the other hand, EGFR mutation in exon 19 is usually presented more in women, in tumors with a smaller maximum diameter, and pleural retraction [
      • Shi Z.
      • Zheng X.
      • Shi R.
      • Song C.
      • Yang R.
      • Zhang Q.
      • Wang X.
      • Lu J.
      • Yu Y.
      • Liu Q.
      • Jiang T.
      Radiological and clinical features associated with epidermal growth factor receptor mutation status of exon 19 and 21 in lung adenocarcinoma.
      ]. This meta-analysis did not assess the differences between EGFR exon mutations due to the marked heterogeneity in the type of alteration reported in the studies.
      The main strengths of our study are that we used a significant sample size of 10355 patients, including only articles with a low or acceptable risk of bias. Also, a sensitivity analysis was performed, providing evidence that all the results presented are robust, and no study is modifying the overall effect by itself. The search strategy was thorough, involving databases worldwide, hand searching, and snowballing methods. Moreover, Publication bias was assessed.
      Some of the limitations of this meta-analysis are that all the articles included were retrospective studies which are more prone to selection bias. Also, machine learning tools interpreted 20.3% of all the CT patterns, while human specialists analyzed the other 79.7%; these differences can lead to inconsistencies in the interpretations. Even though 86% of our studies were confirmed using PCR as the gold standard, it remains another 14% in which other methods were used for confirmation (Immunohistochemistry, FISH, etc.) which can lead to a certain degree of verification bias. All these limitations may affect the internal validity of our study. Nevertheless, these are understandable limitations representing the heterogeneity between studies expected in a meta-analysis. Also, we found publication bias for air bronchogram to detect EGFR mutation, for which a funnel plot was carried out, showing a lack of small studies with negative effects.
      A limitation that may affect the external validity of our findings is that 92% of the sample was from Asia. It has been suggested in previous studies that individuals from Asia have an increased prevalence of EGFR mutation in NSCLC [
      • Midha A.
      • Dearden S.
      • McCormack R.
      EGFR mutation incidence in non-small-cell lung cancer of adenocarcinoma histology: a systematic review and global map by ethnicity (mutMapII).
      ]. Therefore, our study may overestimate the effect of the clinical and CT patterns when extrapolated to other populations. Due to the high number of Asian patients in our meta-analysis, we did not have enough individuals from different continents to perform subgroup analysis according to location. Another limitation is that 97% of all the patients had adenocarcinoma, limiting our findings' generalizability when extrapolated to other histological subtypes of NSCLC. We consider that the results provided in this study are internally valid but must be extrapolated carefully to the non-Asian population and histological subtypes different than adenocarcinoma.

      5. Conclusions

      This meta-analysis indicates that there is enough evidence to conclude that GGO, air bronchogram, spiculated margins, vascular convergence, pleural retraction, early disease stage, non-smoker status, and female gender are significantly associated with EGFR mutation. At the same time, cavitation represents a protective factor for the mutation. The model developed in this study, including all the clinical and CT patterns assessed, showed to be a good predictor for EGFR mutation (AUC: 0.81) and subsequently a determinant for EGFR-TKI response. We consider that the results of this model show strong evidence to encourage the development of clinical scores involving radiological and clinical characteristics to predict EGFR mutation, especially useful in populations in which biopsy cannot be achieved. Further studies evaluating these CT patterns to detect EGFR mutation in individuals different to Asiatic are mandatory to assess if the results in other populations (E.g., Latin, North American, European) correlate with those described in this meta-analysis. The predictive value of these CT patterns in histological subtypes other than adenocarcinoma and the clinical-radiological differences between EGFR mutation in exon 21 and exon 19 remains a mystery, making them a potential source of research.

      Funding

      This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

      CRediT authorship contribution statement

      Andrés Felipe Herrera: Conceptualization, methodology, Formal analysis, Investigation, writing – original draft, writing review and editing, Visualization, Supervision, Project administration. Tatiana Cadavid Camacho: Conceptualization, Supervision, Investigation, Review original draft. Andrés Francisco Vásquez: Investigation, Review – original draft. Valeria del Castillo Herazo: Investigation, Validation. Review – original draft. Juan Guillermo Arámbula: Investigation, Validation, Review – original draft. María Mónica Yepes: Investigation, Review – original draft. Eduard Cadavid Camacho: Investigation, Conceptualization, Review – original draft.

      Declaration of Competing Interest

      The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

      Acknowledgments

      None.

      Appendix A. Supplementary material

      .
      .
      .

      References

        • de Groot P.M.
        • Wu C.C.
        • Carter B.W.
        • Munden R.F.
        The epidemiology of lung cancer.
        Transl. Lung Cancer Res. [Internet]. 2018; 7 (Available from): 220-233https://doi.org/10.21037/tlcr.2018.05.06
        • Sung H.
        • Ferlay J.
        • Siegel R.L.
        • Laversanne M.
        • Soerjomataram I.
        • Jemal A.
        • Bray F.
        BF. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries.
        CA Cancer J. Clin. [Internet]. 2021; 71 (Available from): 209-249https://doi.org/10.3322/caac.21660
        • Herbst R.S.
        • Heymach J.V.
        • Lippman S.M.
        Lung cancer.
        N. Engl. J. Med. [Internet]. 2008; 359 (Available from): 1367-1380https://doi.org/10.1056/NEJMra0802714
        • Avrillon V.
        • Pérol M.
        Alectinib for treatment of ALK-positive non-small-cell lung cancer.
        Future Oncol. [Internet]. 2017; 13 (Available from): 321-335https://doi.org/10.2217/fon-2016-0386
        • Bareschino M.A.
        • Schettino C.
        • Rossi A.
        • Maione P.
        • Sacco P.C.
        • Zeppa R.
        • et al.
        Treatment of advanced non small cell lung cancer.
        J. Thorac. Dis. [Internet]. 2011; 3 (Available from): 122-133https://doi.org/10.3978/j.issn.2072-1439.2010.12.08
        • Lemjabbar-Alaoui H.
        • Hassan O.U.
        • Yang Y.-W.
        • Buchanan P.
        Lung cancer: biology and treatment options.
        Biochim. Biophys. Acta [Internet]. 2015; 1856 (Available from:): 189-210https://doi.org/10.1016/j.bbcan.2015.08.002
        • Shepherd F.A.
        • Rodrigues Pereira J.
        • Ciuleanu T.
        • Tan E.H.
        • Hirsh V.
        • Thongprasert S.
        • Campos D.
        • Maoleekoonpiroj S.
        • Smylie M.
        • Martins R.
        • van Kooten M.
        • Dediu M.
        • Findlay B.
        • Tu D.
        • Johnston D.
        • Bezjak A.
        • Clark G.
        • Santabárbara P.
        • Seymour L.
        Erlotinib in previously treated non-small-cell lung cancer.
        N Engl. J. Med. [Internet]. 2005; 353 (Available from): 123-132https://doi.org/10.1056/NEJMoa050753
        • Yoneda K.
        • Imanishi N.
        • Ichiki Y.
        • Tanaka F.
        Treatment of non-small cell lung cancer with EGFR-mutations.
        J. UOEH [Internet]. 2019; 41 (Available from): 153-163https://doi.org/10.7888/juoeh.41.153
        • Liu Y.
        • Kim J.
        • Qu F.
        • Liu S.
        • Wang H.
        • Balagurunathan Y.
        • Ye Z.
        • Gillies R.J.
        CT features associated with epidermal growth factor receptor mutation status in patients with lung adenocarcinoma.
        Radiology [Internet]. 2016; 280 (Available from): 271-280https://doi.org/10.1148/radiol.2016151455
        • Ramlau R.
        • Krawczyk P.
        • Dziadziuszko R.
        • Chmielewska I.
        • Milanowski J.
        • Olszewski W.
        • Stencel K.
        • Ramlau-Piątek K.
        • Segiet A.
        • Skroński M.
        • Grudny J.
        • Chorostowska-Wynimko J.
        Predictors of EGFR mutation and factors associated with clinical tumor stage at diagnosis: experience of the INSIGHT study in Poland.
        Oncol. Lett. [Internet]. 2017; 14 (Available from): 5611-5618https://doi.org/10.3892/ol.2017.6907
        • Aye P.S.
        • Tin Tin S.
        • McKeage M.J.
        • Khwaounjoo P.
        • Cavadino A.
        • Elwood J.M.
        Development and validation of a predictive model for estimating EGFR mutation probabilities in patients with non-squamous non-small cell lung cancer in New Zealand.
        BMC Cancer [Internet]. 2020; 20 (Available from): 658https://doi.org/10.1186/s12885-020-07162-
        • Lv J.
        • Zhang H.
        • Ma J.
        • Ma Y.
        • Gao G.
        • Song Z.
        • Yang Y.
        Comparison of CT radiogenomic and clinical characteristics between EGFR and KRAS mutations in lung adenocarcinomas.
        Clin. Radiol. [Internet]. 2018; 73 (Available from): 591-598https://doi.org/10.1016/j.crad.2018.01.009
        • Zhang H.
        • Cai W.
        • Wang Y.
        • Liao M.
        • Tian S.
        CT and clinical characteristics that predict risk of EGFR mutation in non-small cell lung cancer: a systematic review and meta-analysis.
        Int. J. Clin. Oncol. [Internet]. 2019; 24 (Available from): 649-659https://doi.org/10.1007/s10147-019-01403-3
        • Cheng Z.
        • Shan F.
        • Yang Y.
        • Shi Y.
        • Zhang Z.
        CT characteristics of non-small cell lung cancer with epidermal growth factor receptor mutation: a systematic review and meta-analysis.
        BMC Med. Imaging [Internet]. 2017; 17 (Available from): 1-10https://doi.org/10.1186/s12880-016-0175-3
        • Mitsudomi T.
        • Yatabe Y.
        Epidermal growth factor receptor in relation to tumor development: EGFR gene and cancer.
        FEBS J. [Internet]. 2010; 277 (Available from): 301-308https://doi.org/10.1111/j.1742-4658.2009.07448.x
        • Herrera Ortiz A.F.
        • Fernández Beaujon L.J.
        • García Villamizar S.Y.
        • Fonseca López F.F.
        Magnetic resonance versus computed tomography for the detection of retroperitoneal lymph node metastasis due to testicular cancer: a systematic literature review.
        Eur. J. Radiol. Open [Internet]. 2021; 8 (Available from)100372https://doi.org/10.1016/j.ejro.2021.100372
        • McGuinness L.A.
        • Higgins J.P.T.
        Risk-of-bias VISualization (robvis): an R package and Shiny web app for visualizing risk-of-bias assessments.
        Res. Syn. Meth. [Internet]. 2020; 1 (Available from): 1-7https://doi.org/10.1002/jrsm.1411
        • Ortiz A.F.H.
        • Camacho E.C.
        • Rojas J.C.
        • Camacho T.C.
        • Guevara S.Z.
        • Cuenca N.T.R.
        A practical guide to perform a systematic literature review and meta-analysis.
        Princ. Pract. Clin. Res. [Internet]. 2021; 7 (Available from): 47-57https://doi.org/10.21801/ppcrj.2021.74.6
        • Hsu J.S.
        • Huang M.S.
        • Chen C.Y.
        • Liu G.C.
        • Liu T.C.
        • Chong I.W.
        • Chou S.H.
        • Yang C.J.
        Correlation between EGFR mutation status and computed tomography features in patients with advanced pulmonary adenocarcinoma.
        J. Thorac. Imaging [Internet]. 2014; 29 (Available from): 357-363https://doi.org/10.1097/RTI.0000000000000116
        • Zhang G.
        • Cao Y.
        • Zhang J.
        • Ren J.
        • Zhao Z.
        • Zhang X.
        • Li S.
        • Deng L.
        • Zhou J.
        Predicting EGFR mutation status in lung adenocarcinoma: development and validation of a computed tomography-based radiomics signature.
        Am J Cancer Res [Internet]. 2021; 11 (Available from): 546-560
        • Qin X.
        • Gu X.
        • Lu Y.
        • Zhou W.
        EGFR-TKI-sensitive mutations in lung carcinomas: are they related to clinical features and CT findings?.
        Cancer Manag. Res. [Internet]. 2018; 10 (Available from): 4019-4027https://doi.org/10.2147/CMAR.S174623
        • Rizzo S.
        • Petrella F.
        • Buscarino V.
        • De Maria F.
        • Raimondi S.
        • Barberis M.
        • Fumagalli C.
        • Spitaleri G.
        • Rampinelli C.
        • De Marinis F.
        • Spaggiari L.
        • Bellomi M.
        CT radiogenomic characterization of EGFR, K-RAS, and ALK mutations in non-small cell lung cancer.
        Eur. Radiol. [Internet]. 2016; 26 (Available from): 32-42https://doi.org/10.1007/s00330-015-3814-0
        • Hasegawa M.
        • Sakai F.
        • Ishikawa R.
        • Kimura F.
        • Ishida H.
        • Kobayashi K.C.T.
        Features of epidermal growth factor receptor-mutated adenocarcinoma of the lung: comparison with nonmutated adenocarcinoma.
        J. Thorac. Oncol. [Internet]. 2016; 11 (Available from): 819-826https://doi.org/10.1016/j.jtho.2016.02.010
        • Mori M.
        • Hayashi H.
        • Fukuda M.
        • Honda S.
        • Kitazaki T.
        • Shigematsu K.
        • Matsuyama N.
        • Otsubo M.
        • Nagayasu T.
        • Hashisako M.
        • Tabata K.
        • Uetani M.
        • Ashizawa K.
        Clinical and computed tomography characteristics of non-small cell lung cancer with ALK gene rearrangement: comparison with EGFR mutation and ALK/EGFR-negative lung cancer.
        Thorac. Cancer [Internet]. 2019; 10 (Available from): 872-879https://doi.org/10.1111/1759-7714.13017
        • Glynn C.
        • Zakowski M.F.
        • Ginsberg M.S.
        Are there imaging characteristics associated with epidermal growth factor receptor and KRAS mutations in patients with adenocarcinoma of the lung with bronchioloalveolar features?.
        J. Thorac. Oncol. [Internet]. 2010; 5 (Available from): 344-348https://doi.org/10.1097/JTO.0b013e3181ce9a7a
        • Dai J.
        • Shi J.
        • Soodeen-Lalloo A.K.
        • Zhang P.
        • Yang Y.
        • Wu C.
        • Jiang S.
        • Jia X.
        • Fei K.
        • Jiang G.
        Air bronchogram: a potential indicator of epidermal growth factor receptor mutation in pulmonary subsolid nodules.
        Lung Cancer [Internet]. 2016; 98 (Available from): 22-28https://doi.org/10.1016/j.lungcan.2016.05.009
        • Zhao J.
        • Dinkel J.
        • Warth A.
        • Penzel R.
        • Reinmuth N.
        • Schnabel P.
        • Muley T.
        • Meister M.
        • Zabeck H.
        • Steins M.
        • Yang J.
        • Zhou Q.
        • Schlemmer H.P.
        • Herth F.J.F.
        • Kauczor H.U.
        • Heussel C.P.
        CT characteristics in pulmonary adenocarcinoma with epidermal growth factor receptor mutation.
        PLoS One [Internet]. 2017; 12 (Available from): 1-14https://doi.org/10.1371/journal.pone.0182741
        • Yang X.
        • Dong X.
        • Wang J.
        • Li W.
        • Gu Z.
        • Gao D.
        • Zhong N.
        • Guan Y.
        Computed tomography‐based radiomics signature: a potential indicator of epidermal growth factor receptor mutation in pulmonary adenocarcinoma appearing as a subsolid nodule.
        Oncologist [Internet]. 2019; 24 (Available from): 1156-1164https://doi.org/10.1634/theoncologist.2018-0706
        • Choi C.M.C.-M.C.-M.
        • Kim M.Y.M.Y.
        • Hwang H.J.H.J.H.J.
        • Lee J.B.J.B.
        • Kim W.S.W.S.
        Advanced adenocarcinoma of the lung: comparison of CT characteristics of patients with anaplastic lymphoma kinase gene rearrangement and those with epidermal growth factor receptor mutation.
        Radiology [Internet]. 2015; 275 (Available from): 272-279https://doi.org/10.1148/radiol.14140848
        • Yano M.
        • Sasaki H.
        • Kobayashi Y.
        • Yukiue H.
        • Haneda H.
        • Suzuki E.
        • Endo K.
        • Kawano O.
        • Hara M.
        • Fujii Y.
        Epidermal growth factor receptor gene mutation and computed tomographic findings in peripheral pulmonary adenocarcinoma.
        J Thorac Oncol [Internet]. 2006; 1 (Available from): 413-416https://doi.org/10.1016/s1556-0864(15)31604-x
        • Suh Y.J.
        • Lee H.J.
        • Kim Y.J.
        • Kim K.G.
        • Kim H.
        • Jeon Y.K.
        • Kim Y.T.
        Computed tomography characteristics of lung adenocarcinomas with epidermal growth factor receptor mutation: a propensity score matching study.
        Lung Cancer [Internet]. 2018; 123 (Available from): 52-59https://doi.org/10.1016/j.lungcan.2018.06.030
        • Tu W.
        • Sun G.
        • Fan L.
        • Wang Y.
        • Xia Y.
        • Guan Y.
        • Li Q.
        • Zhang D.
        • Liu S.
        • Li Z.
        Radiomics signature: a potential and incremental predictor for EGFR mutation status in NSCLC patients, comparison with CT morphology.
        Lung Cancer [Internet]. 2019; 132 (Available from): 28-35https://doi.org/10.1016/j.lungcan.2019.03.025
        • Hong S.J.
        • Kim T.J.
        • Choi Y.W.
        • Park J.S.
        • Chung J.H.L.K.
        Radiogenomic correlation in lung adenocarcinoma with epidermal growth factor receptor mutations: Imaging features.
        Eur. J. Radiol. [Internet]. 2016; 26 (Available from): 3660-3668https://doi.org/10.1007/s00330-015-4196-z
        • Lee Y.
        • Lee H.J.
        • Kim Y.T.
        • Kang C.H.
        • Goo J.M.
        • Park C.M.
        • Paeng J.C.
        • Chung D.H.
        • Jeon Y.K.
        Imaging characteristics of stage I non-small cell lung cancer on CT and FDG-PET: relationship with epidermal growth factor receptor protein expression status and survival.
        Korean J. Radiol. [Internet]. 2013; 14 (Available from): 375-383https://doi.org/10.3348/kjr.2013.14.2.375
        • Yang Y.
        • Yang Y.
        • Zhou X.
        • Song X.
        • Liu M.
        • He W.
        • Wang H.
        • Wu C.
        • Fei K.
        • Jiang G.
        EGFR L858R mutation is associated with lung adenocarcinoma patients with dominant ground-glass opacity.
        Lung Cancer [Internet]. 2015; 87 (Available from): 272-277https://doi.org/10.1016/j.lungcan.2014.12.016
        • Zou J.
        • Lv T.
        • Zhu S.
        • Lu Z.
        • Shen Q.
        • Xia L.
        • Wu J.
        • Song Y.
        • Liu H.
        Computed tomography and clinical features associated with epidermal growth factor receptor mutation status in stage I/II lung adenocarcinoma.
        Thorac. Cancer [Internet]. 2017; 8 (Available from): 260-270https://doi.org/10.1111/1759-7714.12436
        • Hua Wang Matthew B.
        • Schabath Ying
        • Liu Ying
        • Han Qi
        • Li
        • Robert J.
        Gillies et al. Clinical and CT characteristics of surgically resected lung adenocarcinomas harboring ALK rearrangements or EGFR mutations.
        Eur. J. Radiol.[Internet]. 2016; 85 (Available from): 1934-1940https://doi.org/10.1016/j.ejrad.2016.08.023.Clinical
        • Kim T.J.
        • Lee C.T.
        • Jheon S.H.
        • Park J.S.
        • Chung J.H.
        Radiologic characteristics of surgically resected non-small cell lung cancer with ALK rearrangement or EGFR mutations.
        Ann. Thorac. Surg. [Internet]. 2016; 101 (Available from): 473-480https://doi.org/10.1016/j.athoracsur.2015.07.062
        • Park J.
        • Kobayashi Y.
        • Urayama K.Y.
        • Yamaura H.
        • Yatabe Y.
        • Hida T.
        Imaging characteristics of driver mutations in EGFR, KRAS, and ALK among treatment-naïve patients with advanced lung adenocarcinoma.
        PLoS One [Internet]. 2016; 11 (Available from): 1-10https://doi.org/10.1371/journal.pone.0161081
        • Zhou J.Y.
        • Zheng J.
        • Yu Z.F.
        • Xiao W.B.
        • Zhao J.
        • Sun K.
        • Wang B.
        • Chen X.
        • Jiang L.N.
        • Ding W.
        • Zhou J.Y.
        Comparative analysis of clinicoradiologic characteristics of lung adenocarcinomas with ALK rearrangements or EGFR mutations.
        Eur. Radiol. [Internet]. 2015; 25 (Available from): 1257-1266https://doi.org/10.1007/s00330-014-3516-
        • Shi Z.
        • Zheng X.
        • Shi R.
        • Song C.
        • Yang R.
        • Zhang Q.
        • Wang X.
        • Lu J.
        • Yu Y.
        • Jiang T.
        Score for lung adenocarcinoma in China with EGFR mutation of exon 19 Combination of clinical and radiological characteristics analysis.
        Med. (United States) [Internet]. 2018; 97 (Available from)e12537https://doi.org/10.1097/MD.0000000000012537
        • Woo J.H.
        • Kim T.J.
        • Kim T.S.
        • Han J.
        CT features and disease spread patterns in ROS1-rearranged lung adenocarcinomas: comparison with those of EGFR-mutant or ALK-rearranged lung adenocarcinomas.
        Sci. Rep. [Internet]. 2020; 10 (Available from): 16251https://doi.org/10.1038/s41598-020-73533-y
        • Hsu K.H.
        • Chen K.C.
        • Yang T.Y.
        • Yeh Y.C.
        • Chou T.Y.
        • Chen H.Y.
        • Tsai C.R.
        • Chen C.Y.
        • Hsu C.P.
        • Hsia J.Y.
        • Chuang C.Y.
        • Tsai Y.H.
        • Chen K.Y.
        • Huang M.S.
        • Su W.C.
        • Chen Y.M.
        • Hsiung C.A.
        • Chang G.C.
        • Chen C.J.
        • Yang P.C.
        Epidermal growth factor receptor mutation status in stage I lung adenocarcinoma with different image patterns.
        J. Thorac. Oncol. [Internet]. 2011; 6 (Available from): 1066-1072https://doi.org/10.1097/JTO.0b013e31821667b0
        • Shi Z.
        • Zheng X.
        • Shi R.
        • Song C.
        • Yang R.
        • Zhang Q.
        • Wang X.
        • Lu J.
        • Yu Y.
        • Liu Q.
        • Jiang T.
        Radiological and clinical features associated with epidermal growth factor receptor mutation status of exon 19 and 21 in lung adenocarcinoma.
        Sci .Rep. [Internet]. 2017; 7 (Available from): 1-11https://doi.org/10.1038/s41598-017-00511-2
        • Cao Y.
        • Xu H.
        A new predictive scoring system based on clinical data and computed tomography features for diagnosing EGFR-mutated lung adenocarcinoma.
        Curr. Oncol. [Internet]. 2018; 25 (Available from): 132-138https://doi.org/10.3747/co.25.3805
        • Lee H.J.
        • Kim Y.T.
        • Kang C.H.
        • Zhao B.
        • Tan Y.
        • Schwartz L.H.
        • Persigehl T.
        • Jeon Y.K.
        • Chung D.H.
        Epidermal growth factor receptor mutation in lung adenocarcinomas: relationship with CT characteristics and histologic subtypes.
        Radiology [Internet]. 2013; 268 (Available from): 254-264https://doi.org/10.1148/radiol.13112553
        • Sugano M.
        • Shimizu K.
        • Nakano T.
        • Kakegawa S.
        • Miyamae Y.
        • Kaira K.
        • Araki T.
        • Kamiyoshihara M.
        • Kawashima O.
        • Takeyoshi I.
        Correlation between computed tomography findings and epidermal growth factor receptor and KRAS gene mutations in patients with pulmonary adenocarcinoma.
        Oncol. Rep. [Internet]. 2011; 26 (Available from): 1205-1211https://doi.org/10.3892/or.2011.1412
        • Sacconi B.
        • Anzidei M.
        • Leonardi A.
        • Boni F.
        • Saba L.
        • Scipione R.
        • Anile M.
        • Rengo M.
        • Longo F.
        • Bezzi M.
        • Venuta F.
        • Napoli A.
        • Laghi A.
        • Catalano C.
        Analysis of CT features and quantitative texture analysis in patients with lung adenocarcinoma: a correlation with EGFR mutations and survival rates.
        Clin. Radiol. [Internet]. 2017; 72 (Available from): 443-450https://doi.org/10.1016/j.crad.2017.01.015
        • Chen Y.
        • Yang Y.
        • Ma L.
        • Zhu H.
        • Feng T.
        • Jiang S.
        • Wei Y.
        • Wang T.
        • Sun X.
        Prediction of EGFR mutations by conventional CT-features in advanced pulmonary adenocarcinoma.
        Eur. J. Radiol. [Internet]. 2019; 112 (Available from): 44-51https://doi.org/10.1016/j.ejrad.2019.01.005
        • Sabri A.
        • Batool M.
        • Xu Z.
        • Bethune D.
        • Abdolell M.
        • Manos D.
        Predicting EGFR mutation status in lung cancer: proposal for a scoring model using imaging and demographic characteristics.
        Eur. Radiol. [Internet]. 2016; 26 (Available from): 4141-4147https://doi.org/10.1007/s00330-016-4252-3
        • Liu Y.
        • Kim J.
        • Balagurunathan Y.
        • Li Q.
        • Garcia A.L.
        • Stringfield O.
        • Ye Z.
        • Gillies R.J.
        Radiomic features are associated with EGFR mutation status in lung adenocarcinomas.
        Clin. Lung Cancer [Internet]. 2016; 17 (Available from): 441-448.e6https://doi.org/10.1016/j.cllc.2016.02.001
        • Song Z.
        • Zhu H.
        • Guo Z.
        • Wu W.
        • Sun W.
        • Zhang Y.
        Correlation of EGFR mutation and predominant histologic subtype according to the new lung adenocarcinoma classification in Chinese patients.
        Med. Oncol. [Internet]. 2013; 30 (Available from): 645https://doi.org/10.1007/s12032-013-0645-1
        • Zhang Y.
        • Sun Y.
        • Pan Y.
        • Li C.
        • Shen L.
        • Li Y.
        • Luo X.
        • Ye T.
        • Wang R.
        • Hu H.
        • Li H.
        • Wang L.
        • Pao W.
        • Chen H.
        Frequency of driver mutations in lung adenocarcinoma from female never-smokers varies with histologic subtypes and age at diagnosis.
        Clin. Cancer Res. [Internet]. 2012; 18 (Available from): 1947-1953https://doi.org/10.1158/1078-0432.CCR-11-2511
        • Sekine A.
        • Tamura K.
        • Satoh H.
        • Tanaka T.
        • Tsunoda Y.
        • Tanaka T.
        • Takoi H.
        • Lin S.Y.
        • Yatagai Y.
        • Hashizume T.
        • Hayasihara K.
        • Saito T.
        Prevalence of underlying lung disease in smokers with epidermal growth factor receptor-mutant lung cancer.
        Oncol. Rep. [Internet]. 2013; 29 (Available from:): 2005-2010https://doi.org/10.3892/or.2013.2320
        • Moriguchi H.
        • Kim T.Y.
        • Sato C.
        Gefitinib for refractory advanced non-small-cell lung cancer.
        Lancet [Internet]. 2006; 367 (Available from): 299-300https://doi.org/10.1016/S0140-6736(06)68063-X
        • Midha A.
        • Dearden S.
        • McCormack R.
        EGFR mutation incidence in non-small-cell lung cancer of adenocarcinoma histology: a systematic review and global map by ethnicity (mutMapII).
        Am. J. Cancer Res. [Internet]. 2015; 5 (Available from): 2892-2911