ABSTRACT
Non‐small cell lung cancer (NSCLC) patients with EGFR mutations usually receive continuous treatment with tyrosine kinase inhibitors (TKIs) until disease progression. We evaluated whether surgical excision of residual tumor masses offers a survival advantage when performed before progression on TKI therapy. We conducted a retrospective analysis of 230 patients with locally advanced or metastatic EGFR‐mutated NSCLC who were treated with first‐line EGFR TKI. Among these subjects, surgical intervention appeared technically feasible in 57 patients, 41 of whom underwent cytoreductive surgery while 16 were not operated for various reasons. Both surgically treated and nonsurgically treated patients received TKI until disease progression. The median duration of TKI treatment before surgery was 7.3 months. There was no perioperative mortality, while 8/41 (19.5%) patients experienced surgery‐related complications. The median progression‐free survival (PFS) and overall survival (OS) in the surgically treated group were 28.4 months and 46.9 months, respectively. These outcomes were significantly better than in patients with potentially resectable disease who did not undergo surgery (PFS: 15.4 months, p = 0.010; OS: 26.5 months, p = 0.015), or in subjects not amenable to cytoreduction (PFS: 19.0 months, p = 0.006; OS: 31.7 months, p = 0.001). Mutation analysis of residual tumor tissues revealed emerging TKI‐resistant clones in 4 of 21 investigated patients. At the time of data cut‐off, 12/41 (29%) surgically treated patients remained disease‐free. In conclusion, surgical removal of residual tumor masses in EGFR‐mutated NSCLC during response to TKI may confer a survival advantage.
Keywords: activating mutations, EGFR, local treatment, lung cancer, targeted therapy
What's new?
Non‐small cell lung cancer (NSCLC) patients with epidermal growth factor receptor (EGFR) mutations often exhibit significant tumor reduction following tyrosine kinase inhibitor (TKI) therapy. Residual tumors frequently remain, however, leading to drug resistance and disease progression. This study demonstrated that surgical removal of oligopersistent cancer lumps during TKI therapy improves outcomes in EGFR‐driven NSCLC patients when compared to nonoperated subjects. Resistant tumor clones were identified in some residual lesions. The findings suggest that combining TKIs with surgical resection of persistent tumors may delay resistance and improve survival, supporting a multimodal treatment approach.
Non‐small cell lung cancer (NSCLC) patients with epidermal growth factor receptor (EGFR) mutations often exhibit significant tumor reduction following tyrosine kinase inhibitor (TKI) therapy. Residual tumors frequently remain, however, leading to drug resistance and disease progression. This study demonstrated that surgical removal of oligopersistent cancer lumps during TKI therapy improves outcomes in EGFR‐driven NSCLC patients when compared to nonoperated subjects. Resistant tumor clones were identified in some residual lesions. The findings suggest that combining TKIs with surgical resection of persistent tumors may delay resistance and improve survival, supporting a multimodal treatment approach.
Abbreviations
- ABCP
atezolizumab+bevacizumab+paclitaxel+carboplatin
- Atyp res
atypical resection
- Bilob
bilobectomy
- CI
confidence interval
- CR
complete response
- DCR
disease control rate
- ddPCR
droplet digital PCR
- ECOG
Eastern Cooperative Oncology Group
- EGFR
epidermal growth factor receptor
- Ext bilob
extended bilobectomy
- Ext lob
extended lobectomy
- Lob
lobectomy
- MDT
multidisciplinary team
- NA
not assessed
- NR
not reached
- NSCLC
non‐small cell lung cancer
- OR
objective response
- OS
overall survival
- PD
progressive disease
- PFS
progression‐free survival
- PFS2
progression‐free survival on the 2nd line therapy
- Pneum
pneumonectomy
- PR
partial response
- RR
response rate
- TKI
tyrosine kinase inhibitor
- VATS
video‐assisted thoracoscopic surgery
1. Introduction
Non‐small cell lung carcinomas (NSCLCs) carrying EGFR mutations demonstrate pronounced sensitivity to EGFR tyrosine kinase inhibitors (TKIs). The majority of subjects with EGFR‐driven NSCLC respond to gefitinib, erlotinib, afatinib, osimertinib, or other TKIs, as evidenced by significant reduction of tumor lesions. Only a small subset of subjects experience complete response (CR) that is disappearance of visible cancer lesions. The most common result of TKI treatment is partial response (PR). Irrespective of the nature of disease control, EGFR‐mutated NSCLCs are usually treated with continuous TKI administration until disease progression. This treatment ultimately fails because residual tumor cells eventually acquire TKI resistance [1, 2, 3].
Treatment guidelines for several cancer types encourage the sequential use of systemic therapy and local ablation. In this scenario, patients with initially inoperable disease undergo treatment with cytotoxic or targeted drugs, and then residual (oligopersistent) tumor lesions are subjected to eradication by surgical excision or radiotherapy [4]. This sequential application of systemic therapy and surgery is commonly utilized for the management of ovarian cancer, particularly when primary surgical debulking is not feasible due to extensive disease spread or a high risk of perioperative complications [5]. Patients with oligometastatic colorectal cancer and good response to systemic therapy may also be considered for surgical intervention [6]. However, cytoreductive surgery is rarely included in treatment plans for metastatic NSCLC, despite some data suggesting an advantage of this approach in cases of oligometastatic disease [7, 8, 9].
Consideration of local intervention, whether radiological ablation or cytoreductive surgery, may be particularly relevant for patients with near‐complete response to TKI. It is assumed that removal of residual tumor masses allows for elimination of treatment‐resistant cancer clones, which are likely to be the origin of cancer progression after failure of targeted therapy [4]. This approach has been enthusiastically applied in many cancer centers, although the available experience is mainly limited to case series [10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28]. Recent clinical trial demonstrated the promise of radioablation of oligopersistent tumor masses in NSCLC patients receiving osimertinib [29]. This report describes the experience of three cancer centers that systematically considered the option of cytoreductive surgery for subjects with EGFR‐mutated locally advanced or metastatic NSCLC who responded to TKI therapy but showed residual oligopersistent disease.
2. Materials and Methods
2.1. Patients and Procedures
This single‐arm retrospective analysis included patients (n = 230) treated with first‐line TKI therapy for EGFR‐mutated NSCLC between 2015 and 2022. Twenty‐two of these patients underwent cytoreductive surgery during therapy, while 189 subjects remained on therapy without surgical intervention. In addition, we added to the study 19 patients with advanced or metastatic NSCLC who underwent tumor resection during TKI treatment at the City Hospital №62 (Moscow, Russia) within the same time period. The flow chart is presented in Figure 1. The last data cut‐off was performed in October 2025.
FIGURE 1.
Flowchart of patient enrolment.
Surgery was considered for patients with good performance status (0–1) and no significant comorbidities who demonstrated a pronounced response to TKI therapy and had a limited number of potentially operable oligopersistent tumor masses. These patients were discussed by a multidisciplinary team (MDT), which included an experienced thoracic surgeon, a radiotherapist, a medical oncologist, and an anesthesiologist. Surgical removal of oligopersistent disease was not offered to patients who had undergone pneumonectomy or pleurodesis. In addition, subjects with insufficient tumor clearance, for example, persistence of multiple lung metastases, distant lymph node metastases, malignant pleural effusion, or mediastinal invasion during TKI treatment, were not considered for cytoreduction. Therapy with the previously used TKI was restarted within 5–7 days after surgery and continued until disease progression in all included patients.
Progression‐free survival (PFS) and overall survival (OS) were evaluated starting from the first day of TKI treatment until disease progression or death, respectively. PFS2 was evaluated starting from the first day of second‐line treatment until disease progression. Surgical complications were graded according to the Clavien–Dindo classification [30].
2.2. Morphological and Genetic Analysis
Primary EGFR mutational status (ex19del and L858R) was analyzed according to a previously described method [31]. Post‐surgical tissue samples were evaluated for pathologic tumor response according to Hellmann criteria [32]. Wherever available, surgically removed tumor tissues were subjected to molecular genetic analysis. EGFR T790M mutation was detected by droplet digital PCR (ddPCR) for gefitinib‐, erlotinib‐, or afatinib‐treated patients as described in the study [31]; EGFR C797S mutation was analyzed by ddPCR for subjects who received osimertinib. Exome sequencing was carried out according to a previously reported procedure [33].
2.3. Statistical Analysis
We compared treatment outcomes in operated versus nonoperated patients. Furthermore, in order to adjust NSCLC groups for essential confounding parameters, we analyzed available clinical characteristics and CT scans and composed a group of subjects who were potentially amenable to complete or near‐complete cytoreduction before progression on first‐line TKI but were not operated on for various reasons (e.g., preferences of the patients or insufficient confidence of their primary physicians).
PFS, PFS2, and OS were modeled by the Kaplan–Meier method and compared with the log‐rank test. For PFS and OS, univariate Cox proportional hazards analysis was performed for gender, smoking status, stage, EGFR mutation type, disease status (recurrent versus primary metastatic), ECOG performance status, presence of lung, liver, pleura, bone, and adrenal, lymph node, or brain metastases, number of metastatic sites, drug (osimertinib vs. other), and tumor volume decrease. Multivariate Cox proportional hazards analysis was performed to identify significant predictors of PFS and OS. All calculations were conducted using SPSS software (IBM SPSS Statistics v. 20).
We also applied propensity score matching (PSM) to further minimize the influence of confounding factors. The risk of death was estimated with logistic regression for every patient based on the following characteristics: gender, smoking status, T, N, M, stage, EGFR mutation type (ex19del or L858R), ECOG performance status, presence of lung, liver, pleural, bone, adrenal, lymph node, brain, and other metastatic sites, overall number of affected metastatic sites, TKI used in first‐line treatment, objective response, and reduction in tumor volume. Propensity score matching included stratification for brain metastases, mutation type, disease status (primary metastatic versus recurrent), ECOG status, TKI generation (1/2 vs. 3), and age. Operated patients and resectable but nonoperated patients were matched at a 1:2 ratio.
3. Results
3.1. Surgically Treated Patients
Overall, this investigation included 230 patients with inoperable EGFR‐mutated NSCLC who were treated with first‐line TKI. Clinical characteristics of the study subjects are presented in Table 1. Surgical removal of remaining tumor lesions during TKI therapy was performed in 41 of these patients; three of the operated subjects had a complete response (CR) to TKI (patients #201, #209, and #224; Table 2), so their surgery consisted of excision of tissues affected by tumor at the beginning of systemic treatment. Median time to surgery from the start of TKI therapy was 7.3 months (range: 0.7–24.2 months) (Figure 2). The majority of surgically treated patients (37 of 41) received surgery at the peak of tumor response, defined as initial shrinkage of cancer masses followed by stabilization of tumor size upon the next examination; the remaining four subjects were operated on almost immediately after the start of TKI treatment (patients #216, #223, #229, and #230; Table 2). Surgical procedures included sublobar resections (n = 5), lobectomies (n = 16), extended lobectomies (n = 9), bilobectomies (n = 3), and pneumonectomies (n = 3). Three operations involved extrathoracic regions (1 bone resection, 1 splenectomy, and 1 adrenalectomy). Mean duration of surgery was 175 min (range: 60–480 min), and mean blood loss was 214 mL (range: 50–2000 mL) (Table 3).
TABLE 1.
Clinical characteristics of patients included in the study.
| Surgery group A (n = 41) | No surgery group B (n = 16) | Inoperable group C (n = 173) | p | Total | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
| N | % | N | % | N | % | N | % | |||
| Mean age, years (range) | 56.9 (38–77) | 70.1 (53–84) | 66.5 (33–89) | < 0.001 * | 65.1 (33–89) | |||||
| Gender | M | 11 | 26.8% | 1 | 6.3% | 36 | 20.8% | 0.228 | 48 | 20.9% |
| F | 30 | 73.2% | 15 | 93.8% | 137 | 79.2% | 182 | 79.1% | ||
| Smoking | No | 34 | 82.9% | 13 | 81.3% | 136 | 78.6% | 0.936 | 183 | 79.6% |
| Yes | 3 | 7.3% | 1 | 6.3% | 11 | 6.4% | 15 | 6.5% | ||
| Unknown | 4 | 9.8% | 2 | 12.5% | 26 | 15.0% | 32 | 13.9% | ||
| Primary disease status | Recurrent | 9 | 22.0% | 2 | 12.5% | 65 | 37.6% | 0.031 * | 76 | 33.0% |
| Primary metastatic | 32 | 78.0% | 14 | 87.5% | 108 | 62.4% | 154 | 67.0% | ||
| Stage | IIIA | 1 | 2.4% | 0 | 0% | 9 | 5.2% | 0.003 * | 10 | 4.3% |
| IIIB | 6 | 14.6% | 3 | 18.8% | 4 | 2.3% | 13 | 5.7% | ||
| IV | 34 | 82.9% | 13 | 81.3% | 160 | 92.5% | 207 | 90.0% | ||
| Mutation | ex19del | 34 | 82.9% | 11 | 68.8% | 105 | 60.7% | 0.096 | 150 | 65.2% |
| L858R | 6 | 14.6% | 5 | 31.3% | 61 | 35.3% | 72 | 31.3% | ||
| Other | 1 | 2.4% | 0 | 0% | 7 | 4.0% | 8 | 3.5% | ||
| ECOG | 0 | 12 | 29.3% | 0 | 0% | 20 | 11.6% | 0.011 * | 32 | 13.9% |
| 1 | 28 | 68.3% | 14 | 87.5% | 121 | 69.9% | 163 | 70.9% | ||
| 2 | 1 | 2.4% | 2 | 12.5% | 27 | 15.6% | 30 | 13.0% | ||
| 3 | 0 | 0% | 0 | 0% | 5 | 2.9% | 5 | 2.2% | ||
| Lung metastases | Present | 10 | 24.4% | 10 | 62.5% | 109 | 63.0% | 0.000 * | 129 | 56.1% |
| Liver metastases | Present | 1 | 2.4% | 1 | 6.3% | 22 | 12.8% | 0.128 | 24 | 10.5% |
| Pleura metastases | Present | 11 | 26.8% | 6 | 37.5% | 41 | 23.7% | 0.461 | 58 | 25.2% |
| Bone metastases | Present | 13 | 31.7% | 1 | 6.3% | 60 | 34.7% | 0.066 | 74 | 32.2% |
| Adrenal metastases | Present | 2 | 4.9% | 0 | 0% | 6 | 3.5% | 0.665 | 8 | 3.5% |
| Lymph node metastases | Present | 17 | 41.5% | 3 | 18.8% | 39 | 22.5% | 0.036 * | 59 | 25.7% |
| Brain metastases | Present | 8 | 19.5% | 3 | 18.8% | 42 | 24.3% | 0.740 | 53 | 23.0% |
| Other sites | Present | 1 | 2.4% | 2 | 12.5% | 28 | 16.2% | 0.068 | 31 | 13.5% |
| Number of metastatic sites | 1–2 | 32 | 78.0% | 13 | 81.3% | 121 | 69.9% | 0.409 | 166 | 72.2% |
| 3 | 9 | 22.0% | 3 | 18.8% | 52 | 30.1% | 64 | 27.8% | ||
Abbreviations: M, male; F, female.
p values below 0.05 are marked by asterisk and bold font.
TABLE 2.
Individual characteristics of patients who have undergone cytoreductive surgery.
| ID | Age | Gender | Stage | EGFR | ECOG | Drug | 1 L RECIST | Time to surgery | Surgery type | Postoperative complications | Clavien‐Dindo | VATS vs. open | Surgery duration (min) | Blood loss (ml) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 190 | 51 | F | IV | ex19del | 1 | Erlo | PR | 3.8 | Bilob | no | 0 | Open | 150 | 100 |
| 191 | 68 | F | IV | ex19del | 1 | Erlo | PR | 23.4 | Ext lob | no | 0 | Open | 120 | 100 |
| 192 | 68 | F | IV | ex19del | 0 | Gef | PR | 3.8 | Bilob | no | 0 | Open | 160 | 50 |
| 193 | 53 | F | IV | ex19del | 0 | Gef | PR | 6.5 | Ext lob | no | 0 | Open | 140 | 100 |
| 194 | 59 | F | IV | ex19del | 1 | Afa | PR | 21.2 | Bilob | no | 0 | Open | 210 | 100 |
| 195 | 69 | F | IV | ex19del | 1 | Gef | PR | 8.1 | Ext bilob | no | 0 | Open | 205 | 50 |
| 196 | 51 | F | IV | ex19del | 0 | Gef | PR | 5.0 | Ext lob | no | 0 | VATS | 210 | 100 |
| 197 | 59 | F | IV | ex19del | 1 | Gef | PR | 15.3 | Lob | Pneumonia | III | Open | 150 | 100 |
| 198 | 38 | F | IV | L858R | 1 | Erlo | PR | 12.7 | Lob | no | 0 | VATS | 135 | 100 |
| 199 | 55 | F | IV | ex19del | 1 | Gef | PR | 7.3 | Lob | no | 0 | Open | 100 | 100 |
| 200 | 54 | F | IV | ex19del | 0 | Gef | PR | 9.9 | Lob | no | 0 | Open | 200 | 100 |
| 201 | 51 | F | IV | ex19del | 0 | Gef | CR | 23.6 | Lob | Atrial fibrillation | III | Open | 130 | 50 |
| 202 | 61 | M | IV | ex19del | 1 | Erl | PR | 7.8 | Lob | no | 0 | Open | 70 | 50 |
| 203 | 47 | F | IV | L858R | 1 | Osi | PR | 7.2 | Lob | no | 0 | Open | 135 | 50 |
| 204 | 56 | F | IV | ex19del | 1 | Osi | PR | 16.8 | Lob | no | 0 | Open | 100 | 50 |
| 205 | 42 | M | IV | ex19del | 0 | Osi | PR | 9.9 | Lob | no | 0 | VATS | 80 | 100 |
| 206 | 58 | F | IV | ex19del | 0 | Osi | PR | 13.8 | Lob | Atrial fibrillation | III | VATS | 120 | 100 |
| 207 | 59 | F | IV | L858R | 0 | Afa | PR | 8.4 | Lob | no | 0 | VATS | 360 | 200 |
| 208 | 66 | F | IV | ex19del | 1 | Gef | PR | 4.0 | Ext lob | no | 0 | VATS | 160 | 100 |
| 209 | 77 | M | IV | ex19del | 1 | Gef | CR | 4.9 | Splenectomy | no | 0 | VATS | 130 | 150 |
| 210 | 53 | M | IV | ex19del | 1 | Afa | SD | 6.2 | Bone resection | no | 0 | Open | 140 | 100 |
| 211 | 55 | F | IIIB | rare | 1 | Gef | PR | 12.4 | Lob | Chylotorax, pneumonia | III | Open | 240 | 200 |
| 212 | 62 | M | IIIB | ex19del | 1 | Gef | PR | 7.0 | Pneumon | Hematoma | I | Open | 300 | 300 |
| 213 | 59 | M | IV | ex19del | 1 | Gef | PR | 10.5 | Ext lob | no | 0 | Open | 180 | 150 |
| 214 | 69 | F | IV | L858R | 1 | Gef | PR | 8.8 | Ext lob | no | 0 | Open | 180 | 150 |
| 215 | 62 | F | IIIB | ex19del | 0 | Gef | PR | 3.2 | Bilob | Lung vascular embolism | III | Open | 270 | 500 |
| 216 | 43 | F | IIIB | ex19del | 1 | Gef | PR | 0.9 | Pneumon | no | 0 | Open | 200 | 2000 |
| 217 | 56 | F | IV | ex19del | 1 | Gef | PR | 6.3 | Ext lob | no | 0 | Open | 270 | 200 |
| 218 | 59 | F | IIIB | ex19del | 1 | Gef | SD | 7.8 | Pneumon | Intraoperative clinical death | IV | Open | 480 | 800 |
| 219 | 63 | F | IV | ex19del | 1 | Gef | PR | 5.1 | Atyp res | no | 0 | Open | 130 | 50 |
| 220 | 59 | F | IV | ex19del | 2 | Gef | SD | 6.3 | Atyp res | no | 0 | Open | 160 | 100 |
| 221 | 43 | M | IIIB | ex19del | 0 | Gef | PR | 10.0 | Ext lob | no | 0 | Open | 390 | 360 |
| 222 | 62 | M | IV | L858R | 1 | Gef | PR | 4.1 | Atyp res | no | 0 | Open | 60 | 50 |
| 223 | 66 | F | IIIA | ex19del | 1 | Erl | PR | 2.4 | Ext lob | no | 0 | Open | 200 | 150 |
| 224 | 75 | F | IV | ex19del | 1 | Gef | CR | 18.6 | Lob | no | 0 | Open | 160 | 50 |
| 225 | 50 | F | IV | ex19del | 1 | Gef | PR | 16.5 | Atyp res | no | 0 | Open | 60 | 50 |
| 226 | 54 | M | IV | L858R | 0 | Gef | PR | 7.0 | Atyp res | no | 0 | Open | 160 | 50 |
| 227 | 43 | F | IV | ex19del | 1 | Gef | PR | 4.9 | Lob | no | 0 | Open | 180 | 100 |
| 228 | 48 | M | IV | ex19del | 1 | Gef | PR | 3.1 | Lob | Local empiema | III | Open | 270 | 1500 |
| 229 | 45 | F | IV | ex19del | 0 | Afa | SD | 1.4 | Lob | no | 0 | Open | 60 | 50 |
| 230 | 66 | M | IV | ex19del | 1 | Gef | PR | 0.7 | Adrenalectomy | no | 0 | Open | 160 | 50 |
| ID | Pathomorph. response | Viable cells | Molecular analysis of residual tumor | Time from surgery | PFS (mon) | 1st line PD | PD local vs. systemic | 2nd‐line drug | 2nd‐line RECIST | PFS2 (mon) | 2nd‐line PD | OS status | OS (mon) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 190 | IIA | 75% | NSF (d,e) | 13.5 | 17.3 | YES | Distant | Not received | DEAD | 37.6 | |||
| 191 | IIA | 90% | NSF (d) | 7.0 | 30.4 | YES | Both | Not received | DEAD | 68.0 | |||
| 192 | IIB | 5% | NSF (d) | 65.1 | 69.0 | NO | ALIVE | 99.8 | |||||
| 193 | III | 0% | NA | 72.6 | 79.1 | NO | ALIVE | 110.4 | |||||
| 194 | NA | NA | NA | 18.3 | 39.5 | YES | Distant | Not received | DEAD | 42.1 | |||
| 195 | IIB | 7% | NSF (d) | 32.4 | 40.5 | YES | Both | Not received | DEAD | 56.0 | |||
| 196 | IIA | 90% | NSF (d) | 56.8 | 61.8 | NO | DEAD | 61.8 | |||||
| 197 | IIA | 20% | Minor T790M clone (d,e) | 56.3 | 71.6 | NO | ALIVE | 102.1 | |||||
| 198 | NA | NA | NA | 3.0 | 15.7 | YES | Distant | Not received | DEAD | 18.5 | |||
| 199 | IIB | 8% | NSF (d) | 7.8 | 78.6 | YES | Both | Osimertinib | SD | 8.2 | NO | ALIVE | 86.8 |
| 200 | IIA | 30% | Minor T790M clone (d) | 26.9 | 36.7 | YES | Local | Osimertinib | SD | 47.9 | NO | ALIVE | 85.0 |
| 201 | III | 0% | NA | 31.1 | 54.7 | NO | ALIVE | 86.0 | |||||
| 202 | NA | NA | NA | 1.2 | 9.0 | YES | Distant | Not received | DEAD | 34.4 | |||
| 203 | NA | NA | NA | 17.6 | 24.8 | NO | Distant | DEAD | 25.7 | ||||
| 204 | IIA | 90% | NSF (d) | 8.0 | 24.8 | YES | Local | Cisplatin + pemetrexed + bevacizumab + osimertinib | SD | 5.6 | YES | DEAD | 31.7 |
| 205 | IIA | 95% | NSF (d,e) | 2.3 | 12.2 | YES | Local | Atezolizumab + bevacizumab + carboplatin + paclitaxel | SD | 11.6 | YES | DEAD | 38.0 |
| 206 | III | 0% | 10.5 | 24.3 | NO | LOST from follow‐up | 14.1 | ||||||
| 207 | IIA | 60% | NA | 0.4 | 8.8 | YES | Distant | Bevacizumab + carboplatin + paclitaxel | PD | 0.0 | YES | DEAD | 12.3 |
| 208 | IIB | 10% | NA | 31.8 | 35.8 | NO | LOST from follow‐up | 35.8 | |||||
| 209 | III | 0% | NA | 85.1 | 90.1 | YES | Distant | Erlotinib | NA | 0.6 | YES | ALIVE | 93.3 |
| 210 | IIA | 50% | NSF (d,e) | 14.5 | 20.7 | YES | Distant | Osimertinib | SD | 13.8 | YES | DEAD | 47.7 |
| 211 | IIA | 45% | NA | 10.3 | 22.8 | YES | Both | Not received | DEAD | 44.8 | |||
| 212 | NA | NA | NA | 2.1 | 9.2 | YES | Both | Not received | ALIVE | 27.4 | |||
| 213 | IIA | 30% | NA | 14.2 | 24.7 | YES | Distant | Gefitinib | NA | 2.5 | YES | DEAD | 30.9 |
| 214 | IIA | 40% | NSF (d,e) | 44.6 | 53.4 | YES | Distant | Cisplatin + pemetrexed + bevacizumab | SD | 9.8 | NO | ALIVE | 65.6 |
| 215 | NA | NA | NA | 19.8 | 23.1 | YES | Both | Not received | DEAD | 62.1 | |||
| 216 | NA | NA | NA | 19.4 | 20.3 | NO | DEAD | 119.2 | |||||
| 217 | IIB | 5% | Minor T790M clone (d,e) | 6.7 | 13.0 | YES | Distant | Osimertinib | PR | 6.8 | YES | DEAD | 31.4 |
| 218 | NA | NA | NA | 0.0 | 7.9 | NO | DEAD | 7.9 | |||||
| 219 | IIA | 25% | NSF (d,e) | 14.0 | 19.1 | YES | Distant | Osimertinib | PR | 12.8 | YES | DEAD | 36.3 |
| 220 | IIA | 20% | NSF (d,e) | 26.2 | 32.5 | YES | Distant | Not received | DEAD | 39.5 | |||
| 221 | III | 0% | NA | 44.1 | 54.1 | NO | ALIVE | 87.4 | |||||
| 222 | IIA | 75% | NSF (d) | 59.5 | 63.6 | NO | ALIVE | 96.4 | |||||
| 223 | IIA | 30% | NSF (d) | 17.3 | 19.6 | YES | Distant | Gefitinib | SD | 54.5 | NO | ALIVE | 91.4 |
| 224 | NA | NA | NA | 0.4 | 18.1 | YES | Distant | Not received | DEAD | 44.9 | |||
| 225 | IIB | 10% | Disappearance of ex19del (e) | 92.1 | 108.6 | NO | ALIVE | 108.6 | |||||
| 226 | IIB | 5% | NSF (d,e) | 15.2 | 22.2 | YES | Distant | Osimertinib | SD | 46.0 | YES | DEAD | 82.5 |
| 227 | I | 100% | NSF (d,e) | 23.5 | 28.4 | YES | Distant | Not received | DEAD | 46.9 | |||
| 228 | NA | NA | NA | 8.0 | 11.1 | YES | Distant | Not received | DEAD | 14.3 | |||
| 229 | IIB | 10% | NSF (d,e) | 33.0 | 34.4 | YES | Distant | Cisplatin + pemetrexed + bevacizumab | PR | 2.8 | YES | DEAD | 75.9 |
| 230 | IIA | 40% | NA | 7.4 | 8.0 | YES | Distant | Cisplatin + pemetrexed + bevacizumab | PR | 5.8 | YES | DEAD | 35.4 |
Abbreviations: Afa, afatinib; Atyp res, atypical resection; Bilob, bilobectomy; CR, complete response; d, droplet digital PCR; e, exome sequencing; Erl, erlotinib; Ext bilob, extended bilobectomy; Ext lob, extended lobectomy; F, female; Gef, gefitinib; Lob, lobectomy; M, male; NA, not assessed; No PD, no disease progression; NSF, no significant findings upon molecular analysis; OR, objective response; Osi, osimertinib; PD, progressive disease; PFS, progression‐free survival; Pneum, pneumonectomy; PR, partial response; SD, stable disease; VATS, video‐assisted thoracoscopic surgery; vs., versus.
FIGURE 2.
Swimming plot describing pre‐ and post‐surgical history for EGFR‐mutated NSCLC patients subjected to cytoreduction.
TABLE 3.
Surgical outcomes.
| Characteristics | Value |
|---|---|
| Interval from treatment start to surgery. months (min‐max) | 7.0 (1–23.6) |
| Approach | |
| Open | 34 (14.7%) |
| Endoscopic | 7 (20.6) |
| Extent of surgical procedure | |
| Lung atypical resection | 5 (12.2%) |
| Lobectomy | 16 (39.0%) |
| Extended lobectomy | 9 (22.0%) |
| Bilobectomy | 5 (12.2%) |
| Pneumonectomy | 3 (7.3%) |
| Bone resection | 1 (2.4%) |
| Splenectomy | 1 (2.4%) |
| Adrenalectomy | 1 (2.4%) |
| Type of surgery | |
| Complete | 21 (52.2%) |
| Incomplete | 20 (48.8%) |
| Resection | |
| R0 | 33 (80.5%) |
| R1 | 4 (9.8%) |
| R2 | 3 (7.3%) |
| Complications (grade) | 8 (19.5%) |
| I | 1 (2.4%) |
| II | 0 |
| III | 6 (14.6%) |
| IV | 1 (2.4%) |
| Mean duration. minutes (min–max) | 175.5 (60–480) |
| Mean blood loss, ml (min–max) | 214.9 (50–2000) |
Complications were observed in 8/41 (19.5%) of surgically treated patients (Tables 2 and 3). There were six grade III complications according to the Clavien‐Dindo scoring system: two patients with pneumonia and one subject with chylothorax were managed with antibiotics; two patients developed atrial fibrillation, which was managed with antiarrhythmic medications; one patient had pulmonary embolism requiring mechanical ventilation and anticoagulants; the remaining patient experienced local pleural empyema, which was treated with drainage and antibiotics. Grade IV complication was observed in one woman (patient #218): she developed cardiac arrest during the surgical procedure but was successfully resuscitated.
Complete tumor resection, defined as the absence of visible residual macroscopic lesions coupled with the lack of tumor cells in surgical margins, was achieved in 21/41 (51.2%) operated patients. There were no differences in outcomes between patients with complete and near‐complete resection (PFS: 23.1 months (95% CI 17.5–28.7) versus 34.4 months (95% CI 25.5–43.3), p = 0.294; OS: 46.9 months (95% CI 40.8–52.9) versus 44.8 months (95% CI 21.6–68.2), p = 0.591). Complete and major pathologic responses were observed in 5 (16%) and 8 (26%) of 31 evaluable cases, respectively. They were associated with PFS improvement (74.6 months (95% CI 20.6–28.9) versus 24.7 months (95% CI 20.6–28.9), p = 0.005), while this trend did not reach statistical significance for OS (NR (95% CI NR–NR) versus 44.8 months (95% CI 32.9–56.8), p = 0.121).
Residual tumor tissues were available for molecular analysis in 19 subjects treated with first‐ or second‐generation TKIs and in 2 patients treated with osimertinib. Three patients (#197, #200, #217; Table 2) who received either gefitinib or erlotinib showed evidence of emergence of EGFR T790M‐mutated clones, although only a few percent of tumor cells carried this mutation according to ddPCR results and NGS reads (data not shown). Interestingly, the residual tumor sample obtained from patient #225 no longer carried the EGFR ex19del mutation, which was observed in primary NSCLC. This observation cannot be attributed to low content of malignant cells in surgically removed tissue because 14% of NGS reads had the TP53 p.Gly244_Met246del mutation (data not shown). Exome sequencing was performed on 12 residual tumor samples; interestingly, TP53 mutation was detected in only two of these tumors (#205 and #225). Neither of the two residual tumor samples obtained from osimertinib‐treated patients carried the EGFR C797S substitution.
3.2. Survival Analysis in Operated and Nonoperated Patients
Three groups of patients were considered for survival analysis. First, we analyzed surgically treated subjects described above (Group A). Second, the review of clinical records led to the identification of 16 patients who were amenable to surgery during the peak of TKI effect but did not undergo this procedure for various reasons (Group B). Clinical characteristics of these subjects and the proposed extent of surgery are presented in Table S1. The remaining 173 inoperable subjects composed the third group (Group C).
Treatment efficacy analysis is described in Table 4. Median first‐line PFS was significantly higher in operated patients (28.4 months (95% CI 19.3–37.5)) than in potentially operable but nonoperated subjects (15.4 months (95% CI 6.5–24.3); p = 0.010) or cases with inoperable disease (19.0 months (95% CI 16.2–21.7); p = 0.006). It is important to note that initially inoperable stage III patients who achieved operable status apparently derived a greater magnitude of benefit from local treatment compared to subjects with stage IV disease (stage III OS: group A (n = 7) versus group B (n = 3) versus group C (n = 13): 119.2 months (95% CI: NR–NR) versus 46.1 months (95% CI 0–50.9) versus 21.1 months (95% CI 23.4–65.8); A versus B: p = 0.125; A versus C: p = 0.041; B versus C: p = 0.763; stage IV: group A (n = 34) versus group B (n = 13) versus group C (n = 160): 44.9 months (95% CI 8.3–83.9) versus 26.5 months (95% CI 17.2–35.8) versus 31.7 months (95% CI 21.0–42.4); A versus B: p = 0.034; A versus C: p = 0.009; B versus C: p = 0.125).
TABLE 4.
TKI efficacy endpoints.
| Operated (Group A) (n = 41) | Resectable, but nonoperated (Group B) (n = 16) | Unresectable (Group C) (n = 173) | p | Total (n = 230) | |
|---|---|---|---|---|---|
| Drug | |||||
| Erlotinib/gefitinib | 33 (80%) | 14 (88%) | 145 (84%) | 0.010 * | 192 (80%) |
| Afatinib | 4 (10%) | 1 (6%) | 20 (12%) | 25 (11%) | |
| Osimertinib | 4 (10%) | 1 (6%) | 8 (4%) | 13 (10%) | |
| Best response | |||||
| CR | 3 (7%) | 2 (12%) | 7 (4%) | < 0.001 * | 12 (5%) |
| PR | 34 (83%) | 8 (50.0%) | 64 (37%) | 106 (46%) | |
| SD | 4 (10%) | 6 (38%) | 91 (53%) | 101 (44%) | |
| PD | 0 (0%) | 0 (0%) | 8 (5%) | 8 (4%) | |
| Not assessed | 0 (0%) | 0 (0%) | 3 (2%) | 3 (1%) | |
| RR | 37 (90%) | 10 (63%) | 71 (41%) | 0.002 * | 118 (51%) |
| DCR | 41 (100%) | 16 (100%) | 162 (94%) | 0.048 * | 219 (95%) |
|
Median PFS (months) [95% CI] |
28.4 [19.3–37.5] |
15.4 [6.5–24.3] |
19.0 [16.2–21.7] |
A versus B: 0.010 * A versus C: 0.006 * B versus C: 0.443 |
20.7 [18.1–23.3] |
| 19.0 [16.7–21.2] | 0.003 * | ||||
| Progression | 29 (70.7%) | 14 (87.5%) | 123 (71.1%) | 0.365 | 166 (72.2%) |
| PD type | 0.000 * | ||||
| Local | 3 (10.3%) | 10 (71.4%) | 73 (59.3%) | 86 (51.8%) | |
| Systemic | 20 (69%) | 1 (7.1%) | 27 (22%) | 48 (28.9%) | |
| Both | 6 (20.7%) | 0 | 0 | 6 (3.6%) | |
| NA | 0 | 3 (21.4%) | 23 (18.7%) | 26 (15.7%) | |
| Drug | |||||
| Not received | 13 (44.8%) | 4 (28.6%) | 43 (35.0%) | 60 (36.1%) | |
| Erlotinib/gefitinib | 3 (10.3%) | 3 (21.4%) | 6 (4.8%) | 12 (7.2%) | |
| Afatinib | 0 | 2 (14.3%) | 10 (8.1%) | 12 (7.2%) | |
| Osimertinib | 6 (20.7%) | 2 (14.3%) | 39 (31.7%) | 0.164 | 47 (28.3%) |
| Platinum doublet | 6** (20.7) | 3 (21.4%) | 21 (17.1%) | 30 (18.1%) | |
| ABCP | 1 (3.4%) | 0 | 4 (3.3%) | 5 (3.0%) | |
| Best response | |||||
| CR | 0 | 0 | 0 | 0.001 * | 0 |
| PR | 4 (28.6%) | 0 | 11 (13.8%) | 15 (14.4%) | |
| SD | 4 (28.6%) | 5 (50%) | 36 (45%) | 45 (43.3%) | |
| PD | 5 (35.7%) | 0 | 4 (5%) | 9 (8.7%) | |
| Not assessed | 1 (7.1%) | 5 (50%) | 29 (36.3%) | 35 (33.7%) | |
| RR | 4 (28.6%) | 0 | 11 (13.8%) | 15 (14.4%) | |
| DCR | 8 (57.2%) | 5 (50%) | 47 (58.8%) | 60 (57.7%) | |
| Median PFS2 (months) [95% CI] | 12.8 [4.3–21.3] | 2.8 [0.8–4.8] | 6.1 [2.8–9.4] |
A versus B: 0.013 * A versus C: 0.052 B versus C: 0.251 |
6.1 [3.5–8.8] |
| 4.1 [1.3–6.8] | 0.036 * | ||||
| Median OS (months) Stage III |
n = 7 119.2 [NA‐NA] |
n = 3 46.1 [0–50.9] |
n = 13 21.1 [23.4–65.8] |
A versus B: 0.125 A versus C: 0.410 |
46.1 [26.4–65.8] |
| Median OS (months) Stage IV |
n = 34 44.9 [8.3–83.9] |
n = 13 26.5 [17.2–35.8] |
n = 160 21.1 [23.4–65.8] |
A versus B 0.034 A versus C: 0.009 B versus C: 0.125 |
35.7 [28.8–42.6] |
| Median OS (months) [95% CI] | 46.9 [29.9–63.8] | 26.5 [16.8–36.1] | 31.7 [21.8–41.7] |
A versus B: 0.015 * A versus C: 0.001 * B versus C: 0.756 |
36.9 [30.1–43.8] |
| 30.9 [21.5–40.6] | 0.036* | ||||
Abbreviations: Median PFS, median progression‐free survival; Median OS, median overall survival; PR, partial response; CR, complete response; SD, stable disease; PD, disease progression; RR, response rate; DCR, disease control rate.
p values below 0.05 are marked by asterisk and bold font.
1 patient received osimertinib beyond progression in combination with platinum doublet.
Disease progression was observed in 166/230 (72.2%) patients, including 29/41 (70.7%) operated and 137/189 (72.5%) nonoperated subjects. Median duration of systemic treatment from surgery to progression on first‐line TKI was 18.3 months (95% CI 8.4–28.1). There were no differences with regard to this parameter between patients with grade III‐IV surgical complications and subjects without significant perioperative morbidity (19.8 months (95% CI 1.2–38.5) versus 17.6 months (95% CI 7.5–27.8), p = 0.576). There were substantial differences in the pattern of disease progression between operated and nonoperated patients. Among operated subjects, only 3/29 (10.3%) had local failure, 6/29 (20.7%) experienced a combination of regional recurrence and distant metastases, and 20/29 (70.0%) patients presented with distant metastases only. Frequent sites of disease progression involved brain (6/29, 20.7%), lungs and/or pleura (11/29, 37.9%), mediastinal lymph nodes (5/29, 17.2%), distant lymph nodes (4/29, 13.8%), liver (3/29, 10.3%), and adrenal glands (3/29, 10.3%). In the nonoperated group, 83/97 (85.6%) patients with known sites of progression had an increase in the size of the primary tumor.
Information regarding second‐line treatment was available for 166/230 patients, including all subjects in the surgically treated group and 137 nonoperated patients. 60/166 (36.1%) patients did not receive second‐line therapy due to symptomatic progression and rapid deterioration. The most common second‐line therapy was osimertinib (6/29 (20.6%) in group A and 41/137 (29.9%) in groups B + C) followed by platinum doublet with or without atezolizumab (7/29 (24.1%) and 28/137 (20.4%), respectively). The response rates (RR) to second‐line therapy and median PFS2 were significantly better in the operated group (RR: 4/14 (28.6%) versus 11/90 (13.8%), p = 0.001; PFS2: 12.8 months (95% CI 4.3–21.3) versus 4.0 months (95% CI 2.0–6.8), p = 0.036).
Overall survival was longer in the operated group versus nonoperated or nonresectable patients (46.9 months (95% CI 29.9–63.8) versus 26.5 months (95% CI 16.8–36.1) versus 31.7 months (95% CI 21.8–41.7), respectively; A versus B: p = 0.015; A versus C: p = 0.001; B versus C: p = 0.756). A significant difference was also observed when the latter two groups were combined (30.9 months (95% CI 21.5–40.6); p = 0.036).
3.3. Propensity Score Matching Analysis
Univariate Cox proportional hazards analysis conducted for all included patients (n = 230) showed that several factors significantly influenced PFS, including disease status (recurrent versus primary metastatic; p = 0.005), presence of brain metastases at baseline (p = 0.000), number of metastatic sites (0–2 vs. 3+; p = 0.049), decrease in tumor volume (CR + PR vs. SD + PD; p = 0.001), and surgical removal of residual tumor masses (p = 0.003). OS correlated with disease status (recurrent versus primary metastatic; p = 0.001), ECOG status (0–1 vs. 2–3; p < 0.000), presence of bone or brain metastases at baseline (p = 0.001 and p = 0.001, respectively), number of metastatic sites (0–2 vs. 3+; p = 0.000), decrease in tumor volume (CR + PR vs. SD + PD; p = 0.011), and cytoreductive surgery (p = 0.003) (Table 5). NSCLC patients treated with cytoreductive surgery were enriched for subjects with excellent or good ECOG (p = 0.011) and primary metastatic disease (p = 0.031) (Table 1). There were no statistically significant differences between operated and nonoperated patients with regard to other characteristics significant for PFS and OS.
TABLE 5.
Clinical characteristics influence on PFS and OS in univariate analysis.
| Characteristic | No. | Median PFS (95% CI) | p | Median OS (95% CI) | p |
|---|---|---|---|---|---|
| Gender | 0.101 | 0.742 | |||
| F | 182 | 21.5 (18.5–24.3) | 36.9 (29.4–44.7) | ||
| M | 48 | 18.3 (12.3–24.3) | 35.4 (19.2–51.7) | ||
| Smoking | 0.087 | 0.975 | |||
| Nonsmoker | 183 | 20.8 (17.9–23.7) | 39.4 (32.1–46.7) | ||
| Smoker | 15 | 34.1 (4.6–63.6) | 42.5 (38.4–46.6) | ||
| Stage | 0.788 | 0.617 | |||
| Locally advanced (II‐III) | 23 | 19.63 (4.8–34.5) | 46.1 (26.4–65.8) | ||
| Metastatic (IV) | 207 | 20.7 (17.8–23.4) | 35.7 (28.8–42.6) | ||
| Mutation | 0.075 | 0.067 | |||
| Ex19del | 149 | 21.7 (17.9–25.6) | 39.5 (34.6–44.5) | ||
| L858R | 72 | 17.8 (14.4–21.1) | 26.0 (16.3–35.7) | ||
| Disease status | 0.005 * | 0.001 * | |||
| Recurrent | 76 | 28.9 (19.7–38.1) | 47.9 (30.1–65.7) | ||
| Primary metastatic | 152 | 18.5 (16.2–20.7) | 31.7 (23.6–39.9) | ||
| ECOG | 0.216 | 0.000 | |||
| 0/1 | 195 | 20.7 (18.1–23.4) | 41.6 (33.7–49.4) | ||
| 2/3 | 35 | 15.2 (7.3–23.1) | 17.9 (14.5–21.4) | ||
| Lung metastases | 0.883 | 0.242 | |||
| Present | 129 | 20.3 (15.9–25.6) | 34.4 (23.6–45.2) | ||
| Absent | 101 | 20.3 (17.5–23.2) | 39.5 (30.8–48.2) | ||
| Liver metastases | 0.072 | 0.116 | |||
| Present | 24 | 21.5 (18.7–24.3) | 33.3 (13.3–53.3) | ||
| Absent | 206 | 15.9 (10.0–21.7) | 37.6 (30.4–44.7) | ||
| Pleura metastases | 0.945 | 0.350 | |||
| Present | 58 | 20.2 (14.5–23.8) | 29.4 (16.8–42.0) | ||
| Absent | 172 | 20.7 (17.6–23.8) | 39.5 (33.7–45.4) | ||
| Bone metastases | 0.578 | 0.001 * | |||
| Present | 74 | 18.3 (18.9–24.6) | 22.4 (16.3–28.6) | ||
| Absent | 156 | 21.7 (18.9–24.5) | 46.1 (37.3–54.9) | ||
| Adrenal metastases | 0.074 | 0.416 | |||
| Present | 8 | 10.9 (6.7–15.2) | 35.4 (23.9–46.9) | ||
| Absent | 222 | 20.7 (18.2–23.3) | 36.9 (29.8–44.0) | ||
| Lymph node metastases | 0.317 | 0.482 | |||
| Present | 59 | 17.3 (9.3–25.2) | 34.6 (28.9–40.2) | ||
| Absent | 171 | 20.8 (17.9–23.6) | 39.4 (31.6–47.2) | ||
| Other sites | 0.390 | 0.065 | |||
| Present | 31 | 16.5 (11.7–21.3) | 28.6 (18.5–38.7) | ||
| Absent | 199 | 20.7 (17.9–23.5) | 39.2 (32.9–45.6) | ||
| Brain metastases | 0.000 * | 0.001 * | |||
| Present | 53 | 15.6 (11.7–19.5) | 25.2 (17.8–32.6) | ||
| Absent | 177 | 22.8 (19.9–25.7) | 42.1 (35.1–49.1) | ||
| No. of sites | 0.049 * | < 0.000 * | |||
| 0/1/2 | 166 | 22.2 (19.4–25.1) | 42.6 (35.0–50.1) | ||
| 3+ | 64 | 16.5 (13.3–19.8) | 23.7 (18.2–29.1) | ||
| Drug | 0.648 | 0.475 | |||
| Osimertinib | 13 | 24.8 (22.8–26.7) | 36.9 (29.3–44.5) | ||
| Other TKI | 217 | 19.8 (17.6–21.9) | 34.4 (21.0–47.9) | ||
| Tumor volume decrease | 0.001 * | 0.011 * | |||
| Yes (CR + PR) | 118 | 23.1 (19.4–26.7) | 42.5 (34.2–50.7) | ||
| No (SD + PD) | 109 | 16.8 (13.4–20.2) | 29.3 (17.6–41.0) | ||
| Surgery | 0.003 * | 0.001 * | |||
| Yes | 41 | 28.4 (19.3–37.5) | 46.9 (29.9–63.8) | ||
| No | 189 | 19.0 (16.7–21.2) | 30.9 (21.3–40.6) |
Note: Excluded: smoking – no information (n = 33), mutation—other (n = 9).
p values below 0.05 are marked by asterisk and bold font.
Multivariate regression analysis (Table 6) revealed significant differences in hazard ratio of death for age (p = 0.003), disease status (p = 0.038), ECOG (p = 0.001), and presence of brain metastases (p = 0.035). Based on these data, we performed propensity score matching. We composed groups of operated and nonoperated patients with comparable prognostic characteristics (Tables 5 and 6 for multivariate and univariate analysis, respectively). Comparison of clinical characteristics of patients with and without propensity score matching is presented in Table S2. PFS comparison between prognostically balanced groups showed significantly longer PFS for operated versus resectable but nonoperated patients (40.5 months (95% CI NR–NR) versus 15.4 months (95% CI 6.5–24.4), log‐rank p = 0.033). OS was numerically longer in the former versus the latter group (44.9 months (95% CI 17.0–72.9) versus 26.5 months (95% CI 16.8–36.1), log‐rank p = 0.103). Kaplan–Meier curves for PFS and OS are presented in Figure 4.
TABLE 6.
Clinical characteristics influence on OS in multivariate regression analysis.
| Characteristic | HR | 95% CI | p |
|---|---|---|---|
| Age | 1.030 | 1.010–1.051 | 0.003 * |
| Gender | 0.957 | 0.593–1.544 | 0.858 |
| Smoking | 0.825 | 0.422–1.611 | 0.573 |
| Stage | 1.367 | 0.708–2.639 | 0.351 |
| Mutation | 0.720 | 0.488–1.061 | 0.097 |
| Disease status | 0.654 | 0.438–0.976 | 0.038 * |
| ECOG | 0.421 | 0.249–0.709 | 0.001 * |
| Lung metastases | 0.871 | 0.551–1.376 | 0.555 |
| Liver metastases | 0.613 | 0.339–1.105 | 0.104 |
| Pleura metastases | 0.734 | 0.447–1.206 | 0.223 |
| Bone metastases | 0.647 | 0.390–1.074 | 0.093 |
| Adrenal metastases | 1.091 | 0.371–3.203 | 0.874 |
| Lymph node metastases | 1.084 | 0.646–1.818 | 0.759 |
| Other sites | 0.823 | 0.462–1.466 | 0.508 |
| Brain metastases | 1.674 | 1.036–2.702 | 0.035 * |
| No. of sites | 1.068 | 0.488–2.334 | 0.869 |
| Drug | 0.506 | 0.230–1.109 | 0.089 |
| Tumor volume decrease | 0.784 | 0.540–1.136 | 0.198 |
| Surgery | 1.312 | 0.726–2.369 | 0.367 |
p values below 0.05 are marked by asterisk and bold font.
FIGURE 4.
Kaplan–Meier curves of survival without progression (A) and overall survival (B) for groups after propensity score matching in operated (purple—Group A) and operable at nadir, but nonoperated (green—Group B).
4. Discussion
This study revealed that NSCLC patients who underwent cytoreductive surgery for residual tumor masses during EGFR TKI therapy had significantly better disease outcomes compared to the remaining subjects (Table 4, Figure 3). This observation is expected, given that the former group of NSCLC cases is enriched for patients with relatively low tumor burden, good performance status, and pronounced response to TKI treatment. However, we were able to identify subjects who seemed amenable to cytoreduction but did not undergo surgery for various reasons. Strikingly, the benefit of surgery remained evident when the latter category of NSCLC was used for comparison (Table 4, Figure 4). To further reduce the role of confounding parameters, an additional analysis was performed using propensity score matching. This comparison also suggested a role for surgery in improving disease outcomes (Figure 4). These data are in line with treatment attitudes accepted for other cancer types, particularly ovarian and colorectal cancer [5, 6]. Still, the obtained results require validation in randomized prospective trials, especially given that in the presented data set, surgery itself did not retain an independent prognostic role in multivariate regression analysis (Table 6).
FIGURE 3.
Kaplan–Meier curves for progression‐free (A) and overall survival (B) in operated (Group A), operable at nadir, but nonoperated (Group B), and inoperable (Group C).
There are only a few systematic studies involving consolidation strategy for NSCLC patients. Gomez et al. considered patients treated with first‐line chemotherapy or, in case of EGFR mutation or ALK rearrangement, with TKI, and compared outcomes in subjects undergoing surgical or radiological ablation versus subjects receiving conservative treatment [9]. This study was terminated before the planned accrual because of the clear advantage of the former approach (PFS: 11.9 months vs. 3.9 months, p = 0.0054). A similarly designed study, which included only chemotherapy‐treated patients without actionable mutations and involved radiotherapy for residual tumor masses, also stopped early due to significant PFS improvement (9.7 months vs. 3.5 months, p = 0.01) [34]. Promising results were also obtained in a recent clinical trial that applied local radiotherapeutic consolidation to osimertinib‐treated patients [35].
The selection of patients for cytoreductive surgery is always subject to some bias. In our study, this selection was done by a multidisciplinary team, which considered the probability of complete or near‐complete cytoreduction, associated risks, and patient preferences. As a consequence, operated patients were generally younger and more frequently had ECOG status 0 compared to potentially operable but nonoperated subjects (Table 1). It is also essential to acknowledge that some patients with rapid symptomatic improvement underwent surgery during or after the first cycle of TKI therapy (cases #216, #223, #229, and #230), that is, they did not complete a meaningful induction treatment. Overall, our data suggest that a substantial portion of NSCLC patients undergoing TKI treatment may be considered for surgical removal of residual metastatic lesions. Use of different TKIs may also be associated with some variability in disease outcomes. Although third‐generation EGFR TKI osimertinib, either alone or in combination with platinum doublet, is currently a preferable option for upfront treatment of metastatic or locally advanced NSCLC [36, 37], the majority of patients in our study received first‐ or second‐generation inhibitors (Table 2). Indeed, many subjects started treatment before the incorporation of osimertinib into the standards of first‐line NSCLC therapy, and access to this drug remains complicated due to its high cost.
Importantly, there was no surgery‐related mortality. However, the frequency of perioperative morbidity observed in our data set was substantial (37%), given that other studies involving cytoreductive surgery for metastatic NSCLC reported complications in 13%–24% of patients [8, 38, 39, 40, 41, 42]. Direct comparison of these investigations is difficult due to variations in the extent of surgery and patient profile. For example, studies on neoadjuvant TKI treatment demonstrated significantly lower perioperative morbidity (6.5%–15.5%), which is almost certainly attributed to the selection of patients with localized disease [17, 28, 43, 44, 45].
Although the majority of patients subjected to surgery experienced significant clinical reduction of tumor lesions, a major pathologic response was observed in less than half of the subjects, with only 1 in 5 NSCLCs showing a complete pathologic response. This is compatible with the clinical experience of TKI use, as virtually all TKI‐exposed NSCLCs eventually progress due to persistence of TKI‐resistant tumor cell clones [36].
Molecular analysis of tumor tissue revealed only a few instances of emerging TKI‐resistant tumor clones. This is compatible with the fact that surgery was undertaken before NSCLC progression, so we were able to catch only the beginning of the trajectory of tumor escape from therapy. Interestingly, only 2 of 12 informative tumors carried a TP53 mutation. Although TP53 analysis was done only for residual tumor tissues, it is almost certain that the same alteration persisted in the primary NSCLCs. TP53 co‐mutations are observed in up to half of EGFR‐driven tumors and are associated with poor responsiveness to EGFR inhibitors [46]. Low incidence of TP53 mutations in surgically treated NSCLC patients is likely attributed to bias towards subjects with significant TKI‐induced tumor shrinkage.
In conclusion, our study suggests that some patients receiving EGFR TKIs may benefit from cytoreductive surgery. To overcome limitations of this investigation, future studies in this field need to utilize rigorous criteria for patient selection, ensure proper size of the control group, and avoid diversity of EGFR inhibitors by using osimertinib or similar third‐generation drugs.
Author Contributions
Fedor Moiseenko: conceptualization, methodology, formal analysis, investigation, data curation, writing – original draft, writing – review and editing, supervision, project administration. Daniil Stroyakovskiy: investigation, writing – review and editing. Evgeniy Levchenko: investigation, conceptualization. Mikhail Fedyanin: conceptualization, methodology, formal analysis. Albina Gabina: investigation, data curation. Ani Oganesian: investigation, writing – original draft. Sergey Belukhin: investigation. Maria Stepanova: investigation. Ekaterina Stepanova: investigation, data curation. Elizaveta Artemyeva: investigation, data curation. Alexander Myslik: investigation. Nikita Volkov: conceptualization, resources. Alexei Bogdanov: conceptualization, resources. Tatiana Shuginova: investigation. Kirill Kryukov: investigation. Valeria Kheinshtein: investigation. Arif Allakhverdiev: validation. Ivan Zaberegniy: investigation. Yana Akhmadiyarova: investigation. Mikhail Popov: investigation. Anastasiya Danilova: investigation. Dmitriy Sehniaidze: investigation. Nikita Savelov: investigation. Nikita Levchenko: investigation, writing – review and editing. Evgeniy Slugin: investigation. Olga Lopushanskaya: investigation. Ekaterina Kuligina: investigation. Svetlana Aleksakhina: investigation. Aglaya Iyevleva: investigation. Alexandr Ivantsov: investigation. Vladimir Moiseyenko: conceptualization, writing – review and editing, methodology. Evgeny Imyanitov: conceptualization, methodology, writing – review and editing, supervision, project administration.
Funding
This work has been supported by the Russian Science Foundation (grant 24‐45‐10014).
Ethics Statement
The study was approved by the local Ethics Committee (14.03.2023 N3).
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Table S1: Individual characteristics of patients with limited volume of disease and possibility of complete resection selected based on CT data analysis.
Table S2: Clinical characteristics of selected groups after propensity score matching.
Acknowledgements
We are cordially grateful to Dr. Priscilla S. Amankwah and Max Rosenoer for critical reading and editing of this manuscript.
Moiseenko F., Stroyakovskiy D., Levchenko E., et al., “Surgical Removal of Residual Tumor Masses in Patients Undergoing Targeted Therapy for EGFR‐Mutated Locally Advanced or Metastatic Lung Cancer,” International Journal of Cancer 159, no. 9 (2026): 2244–2261, 10.1002/ijc.70543.
[Correction added on 22 July 2026, after first online publication: The copyright line was changed.]
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.
References
- 1. Passiglia F., Bironzo P., Bertaglia V., Listì A., Garbo E., and Scagliotti G. V., “Optimizing the Clinical Management of EGFR‐Mutant Advanced Non‐Small Cell Lung Cancer: A Literature Review,” Translational Lung Cancer Research 11 (2022): 935–949. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Pecci F., Cantini L., Metro G., et al., “Non‐Small‐Cell Lung Cancer: How to Manage EGFR‐Mutated Disease,” Drugs in Context 11 (2022): 1–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Laface C., Maselli F. M., Santoro A. N., et al., “The Resistance to EGFR‐TKIs in Non‐Small Cell Lung Cancer: From Molecular Mechanisms to Clinical Application of New Therapeutic Strategies,” Pharmaceutics 15 (2023): 15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Ravella R., Feliciano E. J. G., Dee E. C., Gomez D. R., and Iyengar P., “Management of Oligoprogressive and Oligopersistent Disease in Advanced NSCLC,” Clinical Advances in Hematology & Oncology 23 (2025): 40–50. [PubMed] [Google Scholar]
- 5. Lheureux S., Gourley C., Vergote I., and Oza A. M., “Epithelial Ovarian Cancer,” Lancet 393 (2019): 1240–1253. [DOI] [PubMed] [Google Scholar]
- 6. Adam R., Kitano Y., Abdelrafee A., Allard M. A., and Baba H., “Debulking Surgery for Colorectal Liver Metastases: Foolish or Chance?,” Surgical Oncology 33 (2020): 266–269. [DOI] [PubMed] [Google Scholar]
- 7. Thai A. A., Solomon B. J., Sequist L. V., Gainor J. F., and Heist R. S., “Lung Cancer,” Lancet 398 (2021): 535–554. [DOI] [PubMed] [Google Scholar]
- 8. Congedo M. T., Cesario A., Lococo F., et al., “Surgery for Oligometastatic Non‐Small Cell Lung Cancer: Long‐Term Results From a Single Center Experience,” Journal of Thoracic and Cardiovascular Surgery 144 (2012): 444–452. [DOI] [PubMed] [Google Scholar]
- 9. Gomez D. R., G. R. Blumenschein, Jr. , Lee J. J., et al., “Local Consolidative Therapy Versus Maintenance Therapy or Observation for Patients With Oligometastatic Non‐Small‐Cell Lung Cancer Without Progression After First‐Line Systemic Therapy: A Multicentre, Randomised, Controlled, Phase 2 Study,” Lancet Oncology 17 (2016): 1672–1682. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Takamochi K., Suzuki K., Sugimura H., et al., “Surgical Resection After Gefitinib Treatment in Patients With Lung Adenocarcinoma Harboring Epidermal Growth Factor Receptor Gene Mutation,” Lung Cancer 58 (2007): 149–155. [DOI] [PubMed] [Google Scholar]
- 11. Kappers I., Klomp H. M., Burgers J. A., Van Zandwijk N., Haas R. L., and van Pel R., “Neoadjuvant (Induction) Erlotinib Response in Stage IIIA Non‐Small‐Cell Lung Cancer,” Journal of Clinical Oncology 26 (2008): 4205–4207. [DOI] [PubMed] [Google Scholar]
- 12. Levchenko E. V., Moiseyenko V. M., Matsko D. E., et al., “Down‐Staging of EGFR Mutation‐Positive Advanced Lung Carcinoma With Gefitinib Followed by Surgical Intervention: Follow‐Up of Two Cases,” Onkologie 32 (2009): 674–677. [DOI] [PubMed] [Google Scholar]
- 13. Hishida T., Nagai K., Mitsudomi T., et al., “Salvage Surgery for Advanced Non‐Small Cell Lung Cancer After Response to Gefitinib,” Journal of Thoracic and Cardiovascular Surgery 140 (2010): e69‐71. [DOI] [PubMed] [Google Scholar]
- 14. Ellerhorst J. A., Greene V. R., Ekmekcioglu S., et al., “Clinical Correlates of NRAS and BRAF Mutations in Primary Human Melanoma,” Clinical Cancer Research 17 (2011): 229–235. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Aukema T. S., Kappers I., Olmos R. A., et al., “Is 18F‐FDG PET/CT Useful for the Early Prediction of Histopathologic Response to Neoadjuvant Erlotinib in Patients With Non‐Small Cell Lung Cancer?,” Journal of Nuclear Medicine 51 (2010): 1344–1348. [DOI] [PubMed] [Google Scholar]
- 16. Zhao P., Zhen H., Zhao H., Zhao L., and Cao B., “Efficacy and Safety of Adjuvant EGFR‐TKIs for Resected Non‐Small Cell Lung Cancer: A Systematic Review and Meta‐Analysis Based on Randomized Control Trials,” BMC Cancer 22 (2022): 328. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Xiong L., Li R., Sun J., et al., “Erlotinib as Neoadjuvant Therapy in Stage IIIA (N2) EGFR Mutation‐Positive Non‐Small Cell Lung Cancer: A Prospective, Single‐Arm, Phase II Study,” Oncologist 24 (2019): 157–164. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Devitt B., Liu W., Salemi R., et al., “Clinical Outcome and Pathological Features Associated With NRAS Mutation in Cutaneous Melanoma,” Pigment Cell & Melanoma Research 24 (2011): 666–672. [DOI] [PubMed] [Google Scholar]
- 19. Ong M., Kwan K., Kamel‐Reid S., and Vincent M., “Neoadjuvant Erlotinib and Surgical Resection of a Stage Iiia Papillary Adenocarcinoma of the Lung With an L861Q Activating EGFR Mutation,” Current Oncology 19 (2012): e222–e226. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Janjigian Y. Y., Park B. J., Zakowski M. F., et al., “Impact on Disease‐Free Survival of Adjuvant Erlotinib or Gefitinib in Patients With Resected Lung Adenocarcinomas That Harbor EGFR Mutations,” Journal of Thoracic Oncology 6 (2011): 569–575. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Funakoshi Y., Takeuchi Y., and Maeda H., “Pneumonectomy After Response to Gefitinib Treatment for Lung Adenocarcinoma,” Asian Cardiovascular & Thoracic Annals 21 (2013): 482–484. [DOI] [PubMed] [Google Scholar]
- 22. Lara‐Guerra H., Chung C. T., Schwock J., et al., “Histopathological and Immunohistochemical Features Associated With Clinical Response to Neoadjuvant Gefitinib Therapy in Early Stage Non‐Small Cell Lung Cancer,” Lung Cancer 76 (2012): 235–241. [DOI] [PubMed] [Google Scholar]
- 23. López‐González A., Almagro E., Salas C., Varela A., and Provencio M., “Use of a Tyrosine Kinase Inhibitor as Neoadjuvant Therapy for Non‐Small Cell Lung Cancer: A Case Report,” Respiratory Medicine Case Reports 9 (2013): 8–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Yamamoto Y., Kodama K., Maniwa T., and Takeda M., “Surgical Resection of Advanced Non‐Small Cell Lung Cancer After a Response to EGFR‐TKI: Presentation of Two Cases and a Literature Review,” Journal of Cardiothoracic Surgery 12 (2017): 98. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Zhai H., Zhong W., Yang X., and Wu Y. L., “Neoadjuvant and Adjuvant Epidermal Growth Factor Receptor Tyrosine Kinase Inhibitor (EGFR‐TKI) Therapy for Lung Cancer,” Translational Lung Cancer Research 4 (2015): 82–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Hishida T., Yoshida J., Aokage K., Nagai K., and Tsuboi M., “Long‐Term Outcome of Surgical Resection for Residual or Regrown Advanced Non‐Small Cell Lung Carcinomas Following EGFR‐TKI Treatment: Report of Four Cases,” General Thoracic and Cardiovascular Surgery 64 (2016): 429–433. [DOI] [PubMed] [Google Scholar]
- 27. Sakanoue I., Hamakawa H., Kaji R., Imai Y., Katakami N., and Takahashi Y., “Sleeve Lobectomy for Lung Adenocarcinoma Treated With Neoadjuvant Afatinib,” Journal of Thoracic Disease 10 (2018): E170–E174. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Zhong W. Z., Chen K. N., Chen C., et al., “Erlotinib Versus Gemcitabine Plus Cisplatin as Neoadjuvant Treatment of Stage IIIA‐N2 EGFR‐Mutant Non‐Small‐Cell Lung Cancer (EMERGING‐CTONG 1103): A Randomized Phase II Study,” Journal of Clinical Oncology 37 (2019): 2235–2245. [DOI] [PubMed] [Google Scholar]
- 29. Sampath S., Rashdan S., Iyengar P., et al., “Osimertinib Plus Consolidative Radiotherapy for Advanced EGFR Mutant Non–small Cell Lung Cancer: A Multicentre, Single‐arm,” Phase 2 Trial eClinicalMedicine 87 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Clavien P. A., Barkun J., de Oliveira M. L., et al., “The Clavien‐Dindo Classification of Surgical Complications: Five‐Year Experience,” Annals of Surgery 25 (2009): 187–196. [DOI] [PubMed] [Google Scholar]
- 31. Lavdovskaia E. D., Iyevleva A. G., Sokolenko A. P., et al., “EGFR T790M Mutation in TKI‐Naive Clinical Samples: Frequency, Tissue Mosaicism, Predictive Value and Awareness on Artifacts,” Oncology Research and Treatment 41 (2018): 634–642. [DOI] [PubMed] [Google Scholar]
- 32. Hellmann M. D., Chaft J. E., W. N. William, Jr. , et al., “Proposal for the Use of Major Pathological Response as a Surrogate Endpoint,” Lancet Oncology 15 (2014): e42‐50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Sokolenko A. P., Bizin I. V., Preobrazhenskaya E. V., et al., “Molecular Profiles of BRCA1‐Associated Ovarian Cancer Treated by Platinum‐Based Therapy: Analysis of Primary, Residual and Relapsed Tumors,” International Journal of Cancer 146 (2020): 1879–1888. [DOI] [PubMed] [Google Scholar]
- 34. Iyengar P., Wardak Z., Gerber D. E., et al., “Consolidative Radiotherapy for Limited Metastatic Non‐Small‐Cell Lung Cancer: A Phase 2 Randomized Clinical Trial,” JAMA Oncology 4 (2018): e173501. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Sampath S., Rashdan S., Iyengar P., et al., “Osimertinib Plus Consolidative Radiotherapy for Advanced EGFR Mutant Non‐Small Cell Lung Cancer: A Multicentre, Single‐Arm, Phase 2 Trial,” EClinicalMedicine 87 (2025): 103435. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Ramalingam S. S., Vansteenkiste J., Planchard D., et al., “Overall Survival With Osimertinib in Untreated, EGFR‐Mutated Advanced NSCLC,” New England Journal of Medicine 382 (2020): 41–50. [DOI] [PubMed] [Google Scholar]
- 37. Jänne P. A., Planchard D., Kobayashi K., et al., “Survival With Osimertinib Plus Chemotherapy in EGFR‐Mutated Advanced NSCLC,” New England Journal of Medicine 394 (2026): 27–38. [DOI] [PubMed] [Google Scholar]
- 38. Ettinghausen S. E. and Burt M. E., “Prospective Evaluation of Unilateral Adrenal Masses in Patients With Operable Non‐Small‐Cell Lung Cancer,” Journal of Clinical Oncology 9 (1991): 1462–1466. [DOI] [PubMed] [Google Scholar]
- 39. Porte H., Siat J., Guibert B., et al., “Resection of Adrenal Metastases From Non‐Small Cell Lung Cancer: A Multicenter Study,” Annals of Thoracic Surgery 71 (2001): 981–985. [DOI] [PubMed] [Google Scholar]
- 40. Zhang C., Wang L., Li W., et al., “Surgical Outcomes of Stage IV Non‐Small Cell Lung Cancer: A Single‐Center Experience,” Journal of Thoracic Disease 11 (2019): 5463–5473. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Granone P., Margaritora S., D'Andrilli A., Cesario A., Kawamukai K., and Meacci E., “Non‐Small Cell Lung Cancer With Single Brain Metastasis: The Role of Surgical Treatment,” European Journal of Cardio‐Thoracic Surgery 20 (2001): 361–366. [DOI] [PubMed] [Google Scholar]
- 42. Bonnette P., Puyo P., Gabriel C., et al., “Surgical Management of Non‐Small Cell Lung Cancer With Synchronous Brain Metastases,” Chest 119 (2001): 1469–1475. [DOI] [PubMed] [Google Scholar]
- 43. Zhang Y., Fu F., Hu H., et al., “Gefitinib as Neoadjuvant Therapy for Resectable Stage II‐IIIA Non‐Small Cell Lung Cancer: A Phase II Study,” Journal of Thoracic and Cardiovascular Surgery 161 (2021): 434–442. [DOI] [PubMed] [Google Scholar]
- 44. Zhong W., Yang X., Yan H., et al., “Phase II Study of Biomarker‐Guided Neoadjuvant Treatment Strategy for IIIA‐N2 Non‐Small Cell Lung Cancer Based on Epidermal Growth Factor Receptor Mutation Status,” Journal of Hematology & Oncology 8 (2015): 54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45. Rizvi N. A., Rusch V., Pao W., et al., “Molecular Characteristics Predict Clinical Outcomes: Prospective Trial Correlating Response to the EGFR Tyrosine Kinase Inhibitor Gefitinib With the Presence of Sensitizing Mutations in the Tyrosine Binding Domain of the EGFR Gene,” Clinical Cancer Research 17 (2011): 3500–3506. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46. Moiseenko F., Kuligina E., Elsakova E., and Imyanitov E., “Epidermal Growth Factor Receptor‐Mutated Lung Carcinomas With Insufficient Response to Epidermal Growth Factor Receptor Inhibitors,” Future Oncology 20 (2024): 2397–2407. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1: Individual characteristics of patients with limited volume of disease and possibility of complete resection selected based on CT data analysis.
Table S2: Clinical characteristics of selected groups after propensity score matching.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.