Clinical efficacies of different neoadjuvant therapies for non-small cell lung cancer

Highlights

• •Comparison of major neoadjuvant therapies for locally advanced NSCLC.
• •Real world cases from one institution using the same evaluation criteria.
• •Disparity between short-term efficacy of radiology and pathology.
• •The impact of MPR on long-term survival for different neoadjuvant therapies.
• •Positive lymph nodes increased the risk of postoperative recurrence.

Abstract

Neoadjuvant therapy followed by surgery is a common clinical strategy for operable non-small cell lung cancer (NSCLC), and the mainstream neoadjuvant therapies include chemoimmunotherapy, targeted therapy, and chemotherapy. However, there is a lack of studies to report the difference in benefits between these treatment modalities in the same institution. Therefore, this study aimed to depict the short-term efficacy of radiology and pathology achieved by different therapies and their impact on long-term survival as well as the underlying clinical significance. A total of 243 NSCLC patients who underwent different neoadjuvant therapies were eligible for inclusion. Demographic, radiological, and pathological features of patients were recorded. The event-free survival (EFS) outcome was analyzed using Kaplan-Meier analysis. The objective response rates (ORR) of primary tumor in the chemoimmunotherapy, targeted therapy, and chemotherapy cohorts were 48.95 %, 57.58 %, and 34.09 % respectively, major pathological response (MPR) rates were 58.74 %, 15.15 %, and 20.83 % (P<.0001), and pathological complete response (pCR) rates were 41.26 %, 0 %, and 11.11 % (P<.0001). For consistency between imaging and pathological evaluation, Cohen's Kappa were 0.275, 0.233, and 0.330. The EFS of MPR group was significantly longer than that of non-MPR group in the chemoimmunotherapy and chemotherapy cohorts (P=.0077**&.0343*, HR=0.3287&0.3715), but this improvement was not observed in the targeted therapy cohort. Neoadjuvant chemoimmunotherapy often underestimates pathological efficacy in imaging but shows consistent long-term outcomes. Neoadjuvant chemotherapy with moderate overall effectiveness has a significant correlation between short-term benefits and reduced recurrence. Neoadjuvant targeted therapy shows remarkable short-term imaging improvements but often fails to convert into sustained long-term survival.

Introduction

The treatment strategy for non-small cell lung cancer (NSCLC) is typically tailored to the patient's specific condition, with surgery being a primary option for early and locally advanced NSCLC. For these patients, according to the 9th edition of tumor-node-metastasis (TNM) staging system, patients with stages I-IIIA (T2–3N1–2a, T4N0–1) disease are considered to have operable NSCLC, and patients with stages IIIA (T1N2b) and IIIB (T2–3N2b, T4N2) are considered to have potentially operable NSCLC. The main goal of surgical treatment is to achieve complete tumor resection, which is suitable for patients with tumors confined to the lungs and no distant metastasis. The 8th edition of the TNM staging system delineates a significant decline in the average 5-year overall survival (OS) for each stage of NSCLC, notably from stages Ib to IIIa, with rates of approximately 68 % for stage Ib, 53–60 % for stage II, and 36 % for stage IIIA [1]. Direct surgery results in a poor prognosis, and thus multidisciplinary treatment (MDT), which involves a team of specialists to optimize patient outcomes, becomes particularly important. The combination of radiotherapy or chemotherapy with surgery during the perioperative period can enhance the overall treatment efficacy. Recent advancements in diagnostic modalities and treatments have led to improved survival rates after the diagnosis of NSCLC [2,3].

The development of neoadjuvant therapy has significantly transformed the treatment landscape for NSCLC. By administering preoperative treatment, the tumor volume can be reduced, which turns initially unresectable tumors into resectable tumors, thereby increasing the success rate of surgical resection and broadening the scope of surgical indications. This approach also allows for an assessment of the tumor's response to treatment before surgery, which informs subsequent postoperative treatments. Furthermore, patients who receive neoadjuvant therapy often experience improved long-term survival rates and a significantly reduced risk of postoperative recurrence. With advances in immunotherapy and targeted therapy, numerous clinical studies on neoadjuvant therapy are actively underway.

Neoadjuvant chemotherapy emerged as a pivotal intervention in the 1990s and has demonstrated good efficacy in reducing tumor burden, achieving downstaging, and markedly increasing surgical resection rates. In 1988, a clinical trial conducted in Spain utilized three courses of chemotherapy (mitomycin, ifosfamide, and cisplatin) as neoadjuvant therapy for stage III lung cancer patients and yielded an impressive objective response rate (ORR) of 73 % and a median survival time of 19 months [4]. Despite its benefits in increasing OS rates compared with surgery alone, neoadjuvant chemotherapy only marginally improves the 5-year OS rate by 5 % [5], which falls short of optimal outcomes.

Since the advent of CheckMate 816, conventional neoadjuvant treatment approaches have shifted from exclusive chemotherapy usage to a combined regimen of immunotherapy and chemotherapy [6]. This integrated chemoimmunotherapy strategy has yielded remarkable outcomes in patients with operable stage IB-IIIA(N2) NSCLC. The immunotherapy group exhibited a pathological complete response (pCR) rate of 24 %, and significantly outperformed the chemotherapy group whose pCR rate was 2.2 %. The median event-free survival (EFS) was 31.6 months in the immunotherapy group compared with 20.8 months in the chemotherapy group [7]. The phase II clinical trial NADIM, which included phase IIIA patients, demonstrated a pCR rate of 63 %, a major pathological response (MPR) rate of 83 %, and a 2-year progression-free survival (PFS) of 77 % [8,9]. The NeoTAP01 study, involving individuals with stage IIIA or T3–4N2 IIIB disease, showed promising results. MPR was achieved in 66.7 % of patients who underwent surgery, pCR was achieved in 50 %, and the 2-year EFS was 67.9 %. Remarkably, the EFS was significantly prolonged in MPR patients compared with non-MPR patients [10]. NEOTORCH, with a pCR rate of 24.8 % and a 2-year EFS rate of 66.7 %, also demonstrated the advantage of an immunotherapy-containing regimen [11]. Similar phase II investigations reported MPR rates in surgical specimens ranging from 60 % to 85 % and pCR rates that ranged from 18 % to 63 % as well as significantly prolonged EFS [[11], [12], [13], [14]]. Moreover, safety profiles remained manageable.

In addition to neoadjuvant chemoimmunotherapy, targeted therapy is another mainstream neoadjuvant treatment modality. Current targeted therapies are primarily used in patients who harbor epidermal growth factor receptor (EGFR) mutations and anaplastic lymphoma kinase (ALK) fusions [15,16]. However, the outcomes of diverse clinical studies exhibit significant variability due to sample size discrepancies. The reported MPR rates ranged from 9.7 % to 67 %, whereas the pCR rates ranged from 0 % to 18.2 %, and the median PFS ranged from 12.1 to 21.5 months [[17], [18], [19]]. Neoadjuvant targeted therapy is still in an exploratory phase and is associated with challenges in optimal treatment duration, surgical timing constraints, and efficacy evaluation metrics. Despite these challenges, neoadjuvant therapy for resectable NSCLC holds substantial promise and offers several potential advantages, including superior tolerability compared with adjuvant therapy, early control of micrometastatic lesions, potential reduction in surgical resection extent leading to better postoperative lung function, and increased rates of complete resection (R0) during surgery [20].

In summary, the various neoadjuvant treatments can provide differing degrees of benefit to patients with operable NSCLC according to existing clinical trial results. However, different centers may report varying pathological scores for patients who receive neoadjuvant therapy, as these scores may be influenced by differences in the evaluation criteria used by each institution [21]. The mechanisms of tumor regression following neoadjuvant therapy across these three treatment modalities remain unclear. Key questions include the following: How do primary tumors and metastatic lymph nodes undergo changes? Can pathological efficacy be visualized by radiographic changes? Do surgical benefits achieved with neoadjuvant therapy contribute to long-term outcomes? Considering that previous studies have focused predominantly on the effects of a single treatment approach and that studies of different treatment approaches from multiple centers may use different evaluation criteria, we performed a retrospective analysis of real-world cases from a single center to systematically compare the similarities and differences among the three primary neoadjuvant treatment strategies. This study aims to provide insights that can improve the diagnosis and treatment of NSCLC.

Methods

Patient selection

This study enrolled all consecutive patients who underwent radical lung cancer resection with lymph node dissection following neoadjuvant treatment at Shanghai Chest Hospital from January 2020 to January 2022. The specific type of surgery was not restricted and included segmentectomy, wedge resection, sleeve lobectomy, and complete lobectomy. All patients were deemed suitable for neoadjuvant therapy after MDT discussion and subsequently underwent the surgery. This comprehensive approach ensures that each patient receives personalized treatment tailored to their specific clinical condition, which aims to maximize the success of surgical outcomes and improve overall prognosis. Neoadjuvant treatments included immunotherapy combined with chemotherapy, targeted therapy alone, and chemotherapy alone. The inclusion criteria necessitated a pathologic diagnosis of NSCLC, an initial clinical stage of operable or potentially operable lung cancer (cT1–4N0–2M0, stage I-IIIB according to the 9th edition of the TNM staging system), satisfactory general health for radical lung cancer resection, and completion of neoadjuvant treatment prior to surgery. The exclusion criteria included diagnosis of pathological types such as small cell lung cancer (SCLC), undifferentiated cancer, the presence of multiple primary lung cancer lesions, the presence of distant tumor metastasis, other concurrent malignancies, CT indications of supraclavicular lymph node metastasis, and the inclusion of other combined drugs in neoadjuvant treatment (e.g., targeted therapy combined with chemotherapy or concurrent chemoradiotherapy).

Information collection

The extracted case information encompassed various domains, including research characteristics (such as neoadjuvant treatment administered, number of cycles utilized, postoperative adjuvant treatment), patient demographics (age, sex, smoking history, comorbidities, recurrence status, survival outcome), and tumor specifics (location, histological subtype, clinical/pathological stage, tumor regression grade, status of lymph node dissection, treatment response, genetic mutations).

Radiological evaluation

Enrolled patients must undergo a minimum of two CT imaging assessments to serve as the baseline for evaluating the response to neoadjuvant treatment: one at the time of initial diagnosis and another following the completion of the final neoadjuvant treatment prior to surgery. The CT slice thickness must not exceed 5 mm, and the interval between these assessments should be no less than 6–8 weeks. The determination of the primary tumor response was based on the analysis of the two CT scans, guided by the Response Evaluation Criteria in Solid Tumors (RECIST 1.1). A complete response (CR) is identified when the target lesion disappears entirely, and the lymph nodes regress to a normal size (short diameter < 10 mm). A partial response (PR) is recognized if the sum of the diameters of the target lesions decreases by ≥ 30 % compared with the baseline. Disease progression (PD) is indicated by a relative increase of ≥ 20 % and an absolute increase of > 5 mm in the sum of target lesion diameters or the appearance of new lesions. Stable disease (SD) is defined when the shrinkage of the target lesion fails to meet the criteria for PR and when the increase does not reach the level of PD [22]. This evaluation was conducted independently by two physicians, and in cases of discordant results, a third physician was consulted for a collaborative assessment. Moreover, to compare the volume of primary tumor more accurately, the gross tumor volume (GTV) for radiotherapy was delineated for all patients.

Pathological evaluation

Upon reviewing all slices containing the tumor bed, a semiquantitative evaluation method was employed to comprehensively assess the percentage of main components within the tumor bed in a comprehensive manner. Presently, three primary components, namely, residual tumor cells, necrosis, and stroma (predominantly fibrous tissue and inflammatory lesions), are recommended for assessment, with the sum of these components totaling 100 % [23,24]. Each component is documented in increments of 10 %, and any component that represents <10 % is recorded as a specific percentage. Following neoadjuvant treatment, reactive lesions may manifest around the tumor. To ensure more accurate tumor bed size measurements, the boundaries of the tumor bed require further refinement by microscopy. PCR is defined as a lack of residual tumor cells that are detected in the tumor bed or lymph nodes following treatment. If the percentage of residual viable tumor cells within the tumor bed after treatment is ≤ 10 %, the response to treatment is deemed an MPR, irrespective of the presence of residual viable tumor cells in the lymph nodes. When pCR is achieved, it is imperative to verify the accurate positioning of the tumor bed in conjunction with pre-treatment images. Those failing to reach an MPR require scoring on the basis of the tumor regression grade (TRG) classification. TRG 1 signifies the absence of residual active tumor cells, equivalent to pCR; TRG 2 indicates residual active tumor cells ≤ 10 %, equivalent to MPR; TRG 3 indicates 10 % < residual active tumor cells ≤ 50 %; and TRG 4 indicates residual active tumor cells > 50 % [21].

Statistical analysis

Categorical and continuous variables were summarized using descriptive statistics. For continuous variables, statistics such as the number of cases, mean, standard deviation, median, minimum, and maximum were utilized for statistical description. Frequencies and percentages were used for the statistical description of categorical variables. Survival information, specifically EFS (the duration from the start of surgery until any certain negative events occurred), was analyzed using the Kaplan-Meier survival analysis method. For patients with confounding factors, inverse probability weighting (IPW) was applied before the survival analysis. Cox proportional hazards models were used to investigate the associations between treatment response (MPR/non-MPR) and disease-related adverse events. Differences in the medians of paired observations were tested using the Wilcoxon signed-rank test. Associations between categorical variables were assessed using Fisher's exact test or the χ² test. A significance level of P < .05 was considered statistically significant. All the statistical analyses and graphical representations were conducted using SPSS 19.0 (IBM), GraphPad Prism 7.0, and R version 3.6.1.

Results

Clinical characteristics

Ultimately, 248 individuals met the eligibility criteria for inclusion in the study (Fig 1). The distribution across treatment groups was as follows: chemoimmunotherapy cohort (CI, N = 143), targeted therapy cohort (T, N = 33), and chemotherapy cohort (C, N = 72). Factors such as sex, age, smoking history, complications, primary tumor position, clinical TNM stage, pathological type, gene mutation status and pathological efficacy were assessed across each cohort (Fig 2). Additionally, all baseline information is shown in Table 1. Noteworthy differences were observed only in sex, pathological type, gene mutation status and MPR status. The observed discrepancy in pathological types within the targeted therapy cohort, predominantly adenocarcinomas with gene mutations among women, aligns with the existing literature [25]. Moreover, the disparity in the MPR rate demonstrates the transformative impact of chemoimmunotherapy compared with other treatments. With respect to the specific gene mutation types in the targeted therapy cohort, 25/33 patients (75.76 %) had EGFR mutations, 6/33 (18.18 %) were positive for the ALK rearrangement, and 2/33 (6.06 %) had other types of mutations. The evolving trends in pre-treatment stage (c), post-treatment stage (y), and pathological stage (yp) of the primary tumor and lymph nodes across the three cohorts revealed varying degrees of downstaging (Fig 3A, B), which highlights the pivotal role of neoadjuvant therapy in diminishing the tumor burden at each stage [26].

Fig. 1.

Flowchart of the study population. NSCLC: non-small cell lung cancer; SCLC: small cell lung cancer; CT: computed tomography.

Fig. 2.

Clinical characteristics of three neoadjuvant cohorts. The distribution of clinical characteristics among all patients, including treatment group, pathological efficacy, sex, age, smoking history, complications, location of the primary tumor, T stage, N stage, clinical stage, pathological type, and gene mutation. CI: chemoimmunotherapy; T: targeted therapy; C: chemotherapy; MPR: major pathologic response; Ad: adenocarcinoma; Sq: squamous carcinoma.

Table 1.

Basic clinical information of patients in each treatment cohort. *** represents P < .001.

Characteristic	CI	T	C	P-value
N(%)	143(57.7)	33(13.3)	72(29.0)
Sex, N(%)				.000***
male	135(94.4)	13(39.4)	59(81.9)
female	8(5.6)	20(60.6)	13(18.1)
Age, N(%)				.902
<65	79(55.2)	19(57.6)	42(58.3)
≥65	64(44.8)	14(42.4)	30(41.7)
Smoke, N(%)				.053
yes	72(50.3)	9(27.3)	35(48.6)
no	71(49.7)	24(72.7)	37(51.4)
Complication, N(%)				.624
yes	49(34.3)	11(33.3)	20(27.8)
no	94(65.7)	22(66.7)	52(72.2)
Position, N(%)				.134
left	52(36.4)	18(54.5)	26(36.1)
right	91(63.6)	15(45.5)	46(63.9)
T, N(%)				.129
1	17(11.9)	7(21.2)	14(19.4)
2	64(44.8)	16(48.5)	25(34.7)
3	26(18.2)	7(21.2)	20(27.8)
4	36(25.2)	3(9.1)	13(18.1)
N, N(%)				.287
0	27(18.9)	9(27.3)	16(22.2)
1	20(14.0)	5(15.1)	4(5.6)
2	96(67.1)	19(57.6)	52(72.2)
TNM stage, N(%)				.602
I	8(5.6)	2(6.1)	6(8.3)
II	24(16.8)	9(27.3)	12(16.7)
III	111(77.6)	22(66.7)	54(75.0)
Type, N(%)				.000***
adeno	32(22.4)	31(93.9)	30(41.7)
squa	95(66.4)	0(0.0)	35(48.6)
other	16(11.2)	2(6.1)	7(9.7)
Gene, N(%)				.000***
positive	10(7.0)	33(100.0)	26(36.1)
negative	106(74.1)	0(0.0)	39(54.2)
MPR, N(%)				.000***
MPR	84(58.7)	5(15.1)	15(20.8)
non-MPR	59(41.3)	28(84.9)	57(79.2)

Fig. 3.

Staging evaluation of three neoadjuvant cohorts. (A)(B) Heatmap illustrating the corresponding changes in T and N stages of the three cohorts at each period. (C) Differences in T and N stage changes among the three cohorts, respectively. 'c' denotes pre-treatment, 'y' denotes post-treatment / pre-operation, and 'yp' denotes surgical pathology. ** represents P < .01, *** represents P < .001, **** represents P < .0001.

Tumor regression

Radiological efficacy assessment

Among the chemoimmunotherapy cohort, 70/143 patients (48.95 %), including 3 who achieved CR and 67 who achieved PR, underwent effective imaging evaluations (CR+PR). The ORR was 19/33 (57.58 %) patients in the targeted therapy cohort, all of whom achieved PR, whereas in the chemotherapy cohort, 28/72 (34.09 %) patients achieved PR (Fig 4A). Although numerical differences were observed in the ORR, no significant difference was found in the imaging response rate among the three cohorts, which may be due to the limited sample size. The change in T stage (cT-yT) before and after neoadjuvant treatment also revealed that a significant reduction from cT to yT occurred in all three cohorts (P<.0001***) (Fig 3C).

Fig. 4.

Evaluation of the therapeutic effects of three treatment methods on imaging and pathology. (A) Assessment and comparison of the treatment response in the three cohorts according to RECIST criteria. (B) Comparison of PCR and MPR in the primary tumor. (C) TRG distribution of the primary tumor across the three cohorts. (D) Relationship between radiological and pathological efficacy in the three cohorts. (E) Arrangement and distribution characteristics of post/pre-GTV in the three cohorts across different TRGs. CR: complete response; PR: partial response; SD: stable disease; PD: progressive disease; TRG: tumor regression grade; post/preGTV: postoperative/preoperative gross tumor volume. * represents P < .05, **** represents P < .0001.

The rate of lymph nodes transitioning from positive to negative [cN(+)-yN(-)] post-treatment was 31/116 (26.72 %) in the chemoimmunotherapy cohort, 13/24 (54.17 %) in the targeted therapy cohort, and 9/56 (16.07 %) in the chemotherapy cohort, respectively (P=.0021**). Additionally, the cN-yN transition in both the chemoimmunotherapy and targeted therapy cohorts significantly decreased (P<.0001**** & P=.0024**), whereas no significant relief in lymph nodes was observed in the chemotherapy cohort (P=.0579) (Fig 3C).

In addition, the changes in the overall clinical stage of the tumor also revealed that various treatment methods can lead to different degrees of downstaging (Table 2). These findings indicate that chemoimmunotherapy and targeted therapy can effectively control both primary tumors and lymph node lesions to reduce the size of primary tumors and to clear tumor-containing metastatic lymph nodes. In contrast, chemotherapy primarily addresses primary tumor lesions. The radiological efficacy ranking among the three methods for primary tumors or lymph nodes is as follows: targeted therapy alone > combined chemoimmunotherapy > chemotherapy alone.

Table 2.

Radiological downstaging of three cohorts before and after neoadjuvant therapy.

	CI (N = 143)		T (N = 33)		C (N = 72)
Stage	at Study Entry	after Treatment	at Study Entry	after Treatment	at Study Entry	after Treatment
	N(%)
0	0	2(1.4)	0	0	0	0
IA	1(0.7)	24(16.8)	2(6.1)	13(39.4)	3(4.2)	9(12.5)
IB	7(4.9)	7(4.9)	0	5(15.2)	3(4.2)	5(6.9)
IIA	6(4.2)	20(14.0)	4(12.1)	4(12.1)	3(4.2)	7(9.7)
IIB	18(12.6)	20(14.0)	5(15.2)	2(6.1)	9(12.5)	7(9.7)
IIIA	35(24.5)	40(28.0)	9(27.3))	8(24.2)	14(19.4)	22(30.6)
IIIB	76(53.1)	30(21.0)	13(39.4)	1(3.0)	40(55.6)	22(30.6)
IV	0	0	0	0	0	0

Pathological efficacy evaluation

The pCR rates of primary lesions in the combined chemoimmunotherapy cohort, targeted therapy cohort, and chemotherapy cohort were 59/143 (41.26 %), 0/33 (0 %), and 8/72 (11.11 %), respectively, which indicates significant discrepancies. Specifically, a comparison between the targeted therapy and chemotherapy revealed a significant difference (P=.0463*), as well as chemoimmunotherapy versus targeted therapy alone and versus chemotherapy alone (both P<.0001****) (Fig 4B). Simultaneously, the MPR rates of the primary tumors in the three cohorts were 84/143 (58.74 %), 5/33 (15.15 %), and 15/72 (20.83 %) in the chemoimmunotherapy, target therapy and chemotherapy cohorts, respectively. Significant disparities were noted between the targeted therapy and chemotherapy cohorts and between the targeted therapy and chemoimmunotherapy cohorts (both P<.0001****) (Fig 4B). Furthermore, a comparison of TRG ratings among the three cohorts revealed average TRG ratings of 2.245, 3.364, and 3.417, respectively. Significantly different TRG scores were observed between the chemoimmunotherapy cohort and the other two cohorts (both P<.0001****) (Fig 4C).

The order of pathological efficacy was neoadjuvant chemoimmunotherapy > chemotherapy > targeted therapy. Following the completion of neoadjuvant treatment completion, the R0 resection rate (negative margins) reached 97.9 %, markedly improving pathological response.

Relationship between pathology and radiology

With respect to the T stage of primary tumors, a significant difference between yT and ypT was observed solely in the chemoimmunotherapy cohort (P=.0003***), while minimal changes were observed in the other two cohorts. Obviously, significant differences in lymph node N stage between yN and ypN were noted in both the chemoimmunotherapy and chemotherapy cohorts (P<.0001**** & P=.0003***), whereas no significant change was observed in the targeted therapy cohort (P=.8558) (Fig 3C).

Furthermore, chi-square analysis was conducted across the three cohorts according to radiological efficacy (CR+PR indicating excellent response/SD+PD indicating poor response) and pathological efficacy (MPR/non-MPR). In the chemoimmunotherapy cohort (χ2= 3.358, P=.0008***), the sensitivity was 60.71 %, the specificity was 67.80 %, the accuracy was 63.64 %, and the kappa coefficient was 0.275. Despite a minor disparity in efficacy between the two evaluation methods (48.95 % vs. 58.74 %), the sensitivity and specificity were relatively low, which indicates poor consistency between the imaging evaluation and the pathological gold standard (Fig 4D). According to the distribution of GTV changes before and after treatment, nearly all patients with post/pre-GTV < 50 % achieved an MPR, whereas nearly half of the patients with post/pre-GTV > 50 % still achieved an MPR (Fig 4E). This further demonstrates that preoperative imaging evaluation criteria are less accurate, which affects the ability of physicians to accurately judge the true therapeutic effect on tumors.

The targeted therapy cohort (χ2=2.084, P=.0372*) demonstrated a sensitivity of 100 %, a specificity of 50 %, an accuracy of 57.58 %, and a Kappa=0.233 (Fig 4D). Regarding the relationship between changes in the GTV of primary lesions and TRG, it was observed that all individuals achieving MPR exhibited relatively effective results in imaging evaluation, whereas half of those with poor pathological effects still displayed positive responses on imaging (Fig 4E).

In contrast, in the chemotherapy group (χ2=3.076, P=.0031**), the sensitivity was 73.33 %, the specificity was 70.18 %, the accuracy was 71 %, and the kappa coefficient was 0.330 (Fig 4D). In terms of the correlation between the GTV distribution and TRG within the chemotherapy group, the TRG1&2 population predominantly achieved a post/pre-GTV ratio of <50 %. Despite the relatively lower overall MPR rate, some cases for which the post/pre-GTV ratios were below 50 % did not achieve MPR, whereas the majority exceeding this threshold belonged to the TRG4 population (Fig 4E).

Long-term outcomes

CI cohort

Among the initial cohort of 143 patients, 27 (18.88 %) were lost to follow-up, and thus comprehensive data from 116 individuals (69 MPR and 47 non-MPR patients) were available for analysis. The mean follow-up, which was calculated from the date of surgery, spanned 24 months (MPR vs. non-MPR=25 vs. 22 months), while the median EFS remained not reached. Notably, disease progression occurred in 10/69 (14.49 %) patients who achieved MPR and 16/47 (34.04 %) patients who did not achieve MPR. The overall average time to recurrence was 16.7 months, with 19.8 months for the MPR group and 14.8 months for the non-MPR group (P=.0070**, HR=0.3287, 95 % CI 0.15–0.74) (Fig 5A). Among these patients, 8 patients who achieved MPR experienced local recurrence, whereas 2 presented with distant metastases (both cerebral). Conversely, among those who did not achieve MPR, 10 experienced local recurrence and 6 presented with distant metastases. Although the chi-square test revealed no significant difference in the distribution of metastatic patterns (P=.4198), a relatively higher incidence of local recurrence was observed (Fig 5D).

Fig. 5.

The occurrence of recurrence and metastasis in three cohorts. (A)(B)(C) Kaplan-Meier survival analysis of EFS in MPR and non-MPR populations of the CI, T, and C cohorts. (D) Pie chart distribution showing no recurrence, local recurrence, and distant metastasis in MPR and non-MPR populations across the three cohorts. Non-M: non-metastasis.

To closely examine the factors contributing to recurrence and metastasis despite the high rate of MPR or even pCR, and given that the relapsed cohort has typically undergone a specified period of postoperative adjuvant chemoimmunotherapy or immunotherapy alone, as established in clinical trial protocols, we stratified the entire patient population on the basis of ypN-negative and ypN-positive status (Table 3). In particular, ypN(+) patients presented a markedly greater likelihood of postoperative local recurrence/distant metastasis than did their ypN(-) counterparts, a trend that was also accentuated within the MPR population (P=.0047** & P=.0298*). In contrast, no significant discrepancy in recurrence rates was detected between lymph node-negative and positive individuals in the non-MPR cohort (P=.2598) (Fig 6A). Additionally, no pattern was observed in the site of recurrence between lymph node-negative and lymph node-positive patients (Fig 6A).

Table 3.

Relationship between Lymph Node Status and Presence or Absence of Metastasis in Each Cohort. * represents P < .05, ** represents P < .01.

Cohort		N Status	M	non-M	P-value
CI (N = 116)	All	ypN(+)	15(37.50)	25(62.50)	.0047**
		ypN(-)	11(14.47)	65(85.53)
	MPR	ypN(+)	5(31.25)	11(68.75)	.0298*
		ypN(-)	5(9.43)	48(90.57)
	non-MPR	ypN(+)	10(41.67)	14(58.33)	.2598
		ypN(-)	6(26.09)	17(73.91)
T (N = 31)	All	ypN(+)	1(8.33)	11(91.67)	.1082
		ypN(-)	7(36.84)	12(63.16)
	MPR	ypN(+)	0(0.00)	0(0.00)	>0.9999
		ypN(-)	2(40.00)	3(60.00)
	non-MPR	ypN(+)	1(8.33)	11(91.67)	.1696
		ypN(-)	5(35.71)	9(64.29)
C (N = 62)	All	ypN(+)	17(56.67)	13(43.33)	.0034**
		ypN(-)	6(18.75)	26(81.25)
	MPR	ypN(+)	1(50.00)	1(50.00)	.2949
		ypN(-)	1(9.09)	10(90.91)
	non-MPR	ypN(+)	16(57.14)	12(42.86)	.0237*
		ypN(-)	5(23.81)	16(76.19)

Fig. 6.

Factors related to postoperative recurrences. (A)(B)(C) The relationship between recurrence and pathological lymph nodes in the overall, MPR, and non-MPR populations. (D) The relationship between later recurrence and gene mutations in the three cohorts, as well as the relationship between recurrence and specific gene mutation types in the T cohort. * represents P < .05, ** represents P < .01.

This phenomenon emphasizes the intricate interplay between primary tumor control and residual lymph node status in dictating recurrence outcomes. While MPR associated primary tumor control may mitigate the risk of recurrence, the presence of residual positive lymph nodes portends increased tumor burden and subsequent recurrence. Conversely, non-MPR patients with inadequately controlled primary tumors may harbor an existing tumor burden, which may diminish the impact of lymph node status. The higher recurrence rate observed in ypN(+) patients, which is attributed to suboptimal lymph node removal during surgery, highlights the latent risk posed by overlooked lymph nodes as potential metastatic reservoirs.

Among the subset of 116 patients with available gene mutation data, 7/12 (58.33 %) individuals with identified mutations experienced recurrence, whereas 19/104 (18.27 %) patients without mutations experienced recurrence (P=.0048**, RR=3.193) (Fig 6D). While genetic mutations have no discernible impact on immunotherapeutic efficacy during the neoadjuvant phase, they have emerged as formidable prognostic indicators of prolonged patient survival.

T cohort

Among the 31 patients with follow-up data (MPR vs. non-MPR = 5 vs. 26), the overall average follow-up time was 29.8 months (MPR vs. non-MPR = 29.2 months vs. 30 months), and the median EFS was not reached in either the MPR or the non-MPR group. The overall average time to recurrence was 23 months (21.5 months in the MPR group and 23 months in the non-MPR group). Recurrence occurred in 2 of 5 patients (40 %) in the MPR group and in 6 of 26 patients (23.08 %) in the non-MPR group (P=.4577, HR=2.068, 95 % CI 0.30–14.07) (Fig 5B). These data suggest that the prognosis of patients who achieve MPR may not necessarily be better than that of patients who do not achieve MPR.

Given the minimal role of pathological remission of the primary tumor after neoadjuvant induction on long-term patient survival, we further stratified the population with recurrence according to postoperative pathological N stage (Table 3). Only 1 out of 8 patients was lymph node-positive (N1), whereas the remaining 7 patients were N0. We then conducted correlation analyses between lymph node status and recurrence in the MPR and non-MPR groups, and found no difference (Fig 6B). This result eliminates the potential impact of lymph nodes on postoperative tumor recurrence and metastasis. It can be inferred that pathological outcomes following neoadjuvant targeted therapy are weakly correlated with patient prognosis, regardless of primary tumor or lymph node status. Therefore, we speculate that neoadjuvant targeted therapy provides limited long-term benefits for patients with NSCLC and genetic mutations, and that the existing neoadjuvant pathological evaluation indicators are inadequate for predicting long-term patient survival.

An examination of the sites of tumor recurrence revealed 1 local recurrence and 1 distant metastasis in the MPR group, whereas 2 local recurrences and 4 distant metastases were observed in the non-MPR group (Fig 5D). No clear distinction was observed in the recurrence pattern, whether it was local recurrence or distant metastasis.

A comparison of the recurrence rates in patients with different gene mutations revealed that patients with EGFR mutations had a much greater recurrence rate (26.09 %) than those with non-EGFR mutations (12.50 %) (P=.6417, OR=2.471) (Fig 6D). These findings confirm that patients with NSCLC and EGFR mutations have worse long-term survival, even with neoadjuvant targeted therapy before surgery.

C cohort

Excluding the 10 patients (13.89 %) who were lost to follow-up, a total of 62 patients with survival information were analyzed (13 in the MPR group and 49 in the non-MPR group). The overall average follow-up time was 26.1 months (MPR vs. non-MPR = 26.6 months vs. 25.3 months). The overall average time to recurrence was 20.6 months (23.5 months in the MPR group and 20.3 months in the non-MPR group). Recurrence occurred in 2 of 13 patients (15.38 %) in the MPR group, but in 21 of 49 patients (42.86 %) in the non-MPR group (P=.0343*, HR=0.3715, 95 % CI 0.15–0.93) (Fig 5C). The median EFS was not reached in the MPR group, whereas it was 32 months in the non-MPR group. These data indicate that patients whose primary tumors achieve MPR after neoadjuvant chemotherapy have a significantly reduced risk of recurrence and a better long-term prognosis.

Regarding the recurrence pattern, only 1 case of recurrence in the MPR group was distant metastasis, whereas 11 cases of local recurrence and 10 cases of distant metastasis were observed in the non-MPR group. No significant difference was noted in the metastasis site (P=.3061) (Fig 5D).

When the recurrence rates were analyzed according to the pathological lymph node status, the recurrence rate in N(-) patients was 6 of 32 patients (18.75 %), whereas that in N(+) patients was 17 of 30 patients (56.67 %) (P=.0034**, OR=0.1765). No association between ypN stage and relapse was observed in the MPR population (Table 3). However, in the non-MPR group, the recurrence rate in N(-) patients was 5 of 21 patients (23.81 %) and that in N(+) patients was 16 of 28 patients (57.14 %) (P=.0237*, OR=0.2344) (Fig 6C). This finding illustrates that the presence of positive lymph nodes during surgery after neoadjuvant chemotherapy significantly increases the risk of recurrence and metastasis. This result is consistent with previous findings and highlights the need for adjuvant chemoradiotherapy to improve long-term survival. In summary, the inability of the primary tumor to achieve MPR and the presence of uncleared lymph nodes after chemotherapy induction collectively contribute to tumor recurrence, with an average overall recurrence time of more than 2 years.

We also compared recurrence rates among patients with and without gene mutations. It was found that 9 of 22 patients (40.91 %) with gene mutations and 12 of 35 patients (34.29 %) without mutations experienced recurrence after surgery (P=.7786, OR=1.327) (Fig 6D). These findings suggest a weak relationship between gene mutations and recurrence in the neoadjuvant chemotherapy group.

Discussion

The objective of this study was to evaluate the current status of various neoadjuvant therapies by focusing on their respective advantages and limitations. The mainstream treatments include chemoimmunotherapy, targeted therapy, and chemotherapy [20]. This study systematically compared the radiological and pathological efficacy, along with long-term survival outcomes, of these three primary neoadjuvant approaches and assessed the concordance among these results. We hope that this research will improve the evaluation of tumors by clinicians and provide guidance for the comprehensive treatment of operable NSCLC. The study population included stage III patients, which is consistent with existing clinical trials [8,9,11]. Since most studies have focused on the tumor control benefits of individual treatment modalities, comparisons between different treatment methods are lacking. This article examines the short- and long-term efficacy of three mainstream neoadjuvant therapies. Given that previous clinical studies originate from multiple centers, which may introduce subjective evaluation bias, the findings of our single-center retrospective study affirm variations between various neoadjuvant treatment modalities [20].

Preliminary data demonstrate that, compared with previous treatment modalities, neoadjuvant chemoimmunotherapy has unparalleled pathological efficacy. Both the MPR and pCR rates may be exceptionally high; however, disparities exist between imaging assessments and actual pathological responses. Although CT reveals a dense solid mass, only minimal residual tumor cells or even complete absence of tumor cells upon pathological examination can be visualized. Imaging evaluation underestimates the actual efficacy of neoadjuvant chemoimmunotherapy. The difficulty in visualizing pathological response has added obstacles to the selection of treatment after neoadjuvant chemoimmunotherapy.

Compared with chemoimmunotherapy, neoadjuvant targeted therapy may result in lower MPR or pCR rates. However, other efficacy indicators, such as R0 resection and downstaging, remain comparable. In particular, the imaging-based ORR of targeted therapy significantly surpasses that of other treatments and often correlates with the abundance of mutations identified by genetic testing. This can be rationalized by the mechanism of action of targeted drugs, which achieve therapeutic effects by directly inhibiting tumor cell growth, promoting apoptosis, and impeding spread [27]. In this study, although the tumor volume significantly decreased after treatment, the internal tumor cell concentration did not proportionally decrease. As a result, the relative tumor cell content increased, which led to a greater TRG in the targeted therapy cohort than in the other two cohorts. As targeted drugs function primarily through inhibition rather than elimination, achieving pCR is rare.

Conversely, traditional chemotherapy at the neoadjuvant stage offers limited benefits to patients, with suboptimal efficacy observed in both imaging and pathology assessments. Nonetheless, a correlation between the two evaluation methods persists, as radiologically effective responders often exhibit lower pathological TRG. These findings align with the mechanism of action, which involves the destruction of tumor cells and inhibition of their growth and proliferation.

After defining the short-term efficacy indicators of neoadjuvant therapy, it is essential to focus on long-term survival indicators. As both MPR and pCR are merely short-term surrogate markers, the goal of neoadjuvant therapy is to prolong survival and improve the prognosis of patients with NSCLC. Existing clinical studies focus on the survival of patients receiving neoadjuvant chemoimmunotherapy compared with that of patients receiving neoadjuvant chemotherapy, which can prolong EFS and OS [28]. The 2-year EFS rate was also significantly higher in the neoadjuvant treatment groups than in the groups without neoadjuvant treatment [29,30]. Additionally, we found that the patients in the chemoimmunotherapy and chemotherapy cohorts who achieved MPR had a longer EFS. Although neoadjuvant treatment effectively controlled the primary tumor to a significant extent, the presence of positive lymph nodes after neoadjuvant treatment increased the risk of recurrence associated with surgical procedures. In patients with MPR of the primary tumor, the presence of positive lymph nodes after surgery increases tumor mutational burden (TMB) and the risk of recurrence. However, in non-MPR patients, where the primary lesion already exhibits a certain degree of TMB, the impact of positive lymph nodes is relatively diminished. This phenomenon raises important considerations regarding the choice of postoperative adjuvant treatment. For the targeted therapy cohort, not only were the pathological evaluation indicators suboptimal, but the long-term survival outcomes were also poor. In contrast to the other two cohorts, achieving MPR after targeted therapy may be more closely related to changes in the primary tumor volume rather than the number of remaining viable tumor cells. This finding aligns with the mechanism of action of targeted therapy [27,31]. MPR does not necessarily indicate a lower tumor burden, and thus may not translate into a long-term survival benefit.

As a retrospective study, this article possesses inherent limitations. First, the study population was derived from a subset of surgical patients, which may have biased the treatment efficacy compared with that in a broader population. This introduces a degree of information bias, which may impact the generalizability of the findings. Moreover, the sample size of this study is not large enough, especially the targeted therapy cohort, which is also limited by the small number of patients who adopt this treatment model in the real world. As observed, only five individuals in the T cohort achieved MPR. Consequently, the long-term survival data for this group may not be entirely reliable and requires validation through studies with larger sample sizes. Finally, although we selected EFS as the indicator of the long-term efficacy of neoadjuvant therapy due to the short follow-up time, we will also focus on OS in the future. More study endpoints need to be developed to target specific neoadjuvant treatment modalities. Such investigations will contribute significantly to the advancement of clinical practice and treatment strategies in the field.

Our data underscore both the value and limitations of current NSCLC treatments. By systematically assessing the concordance between imaging and pathological evaluations, we challenge established paradigms and reveal the inadequacies of preoperative imaging in accurately gauging therapeutic outcomes; we also advocate for more precise evaluation criteria. Emerging tumor biomarkers, such as minimal residual disease (MRD) and circulating tumor DNA (ctDNA), hold promise for assessing treatment response [32,33]. Additionally, the MPR rate following neoadjuvant therapy needs improvement, particularly for targeted therapies. Combining neoadjuvant targeted therapy with other treatments may lead to better outcomes, but this requires validation in large-scale prospective clinical trials [34,35]. Current imaging and pathological evaluation criteria are insufficient to determine whether patients can benefit from neoadjuvant therapy in specific scenarios. Future research should focus on accurately predicting short-term efficacy after neoadjuvant therapy to guide subsequent treatment decisions [36,37].

In summary, this retrospective clinical study provides an objective comparative analysis of short-term and long-term efficacy across different neoadjuvant treatment modalities for operable NSCLC. The ranking of short-term efficacy was as follows: neoadjuvant chemoimmunotherapy > targeted therapy > chemotherapy. Both chemoimmunotherapy and chemotherapy can convert their MPR into long-term prognostic benefits. Notably, this research revealed critical insights into the prognostic implications of post-treatment lymph node involvement by stratifying patients according to ypN status. Positive lymph nodes after neoadjuvant chemoimmunotherapy and chemotherapy increase the risk of postoperative recurrence. This comprehensive approach not only broadens our understanding of neoadjuvant treatment efficacy but also paves the way for optimizing clinical decision-making and enhancing personalized treatment planning in patients with NSCLC.

Funding information

This work was supported by the Major Research Plan of the National Natural Science Foundation of China (Grant No. 92059206) and the National Natural Science Foundation of China (Grant No. 82103010).

Ethics statement

All procedures in this study Adhered to the Declaration of Helsinki (as revised in 2013). The study was approved by the Institutional Ethical Committee of Shanghai Chest Hospital (No. KS21002). Since this was a retrospective study, the necessity for written informed consent from patients was waived.

CRediT authorship contribution statement

Xin-chen Tan: Writing – original draft, Resources, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Xin-yun Song: Methodology, Investigation. Meng-qi Jiang: Investigation, Conceptualization. Neng-yang Wang: Methodology, Investigation. Jun Liu: Methodology, Investigation. Wen Yu: Methodology, Conceptualization. Qin Zhang: Methodology, Conceptualization. Xu-wei Cai: Investigation, Conceptualization. Wen Feng: Writing – review & editing, Validation, Supervision, Methodology, Funding acquisition. Xiao-long Fu: Writing – review & editing, Supervision, Project administration, Funding acquisition.

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.