Adjuvant chemotherapy appears less effective in undernourished lung cancer patients as assessed by the controlling nutritional status score: a retrospective cohort study.

Introduction

Background
Non-small cell lung cancer (NSCLC) remains one of the leading causes of cancer-related mortality worldwide. Surgery combined with perioperative systemic therapy is a cornerstone of radical treatment for resectable advanced NSCLC. Nevertheless, the 5-year overall survival (OS) rates for patients with pathological stage IB, IIA, IIB, and IIIA disease are approximately 73%, 65%, 56%, and 41%, respectively (1), even after complete resection, highlighting the need to improve systemic control of minimal residual disease.
Postoperative platinum-based adjuvant chemotherapy has been established as a standard of care based on meta-analyses and pooled analyses of randomized trials (2,3), which demonstrate a modest but clinically meaningful 5–15% improvement in 5-year OS in appropriately selected patients (2-4). Recently, immune checkpoint inhibitors (ICIs) and molecular targeted agents such as epidermal growth factor receptor tyrosine kinase inhibitors (EGFR-TKIs) have further improved outcomes in advanced disease and have increasingly been applied in perioperative settings, often in combination with chemotherapy (5-7). However, pivotal trials typically enroll patients with good performance status (PS) and preserved organ function, and may not fully represent the frail or undernourished populations frequently encountered in routine practice.

Rationale and knowledge gap
A substantial proportion of patients with NSCLC are elderly and/or undernourished. Malnutrition and systemic inflammation are recognized prognostic factors in advanced lung cancer, and several composite indices—such as the C-reactive protein-albumin ratio (CAR), neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), prognostic nutritional index (PNI), and controlling nutritional status (CONUT) score—have been associated with outcomes in various malignancies, including lung cancer (8-11). However, it remains unclear whether these indices can predict the magnitude of benefit from platinum-based adjuvant chemotherapy in resectable NSCLC.
In daily clinical practice, the decision to administer adjuvant chemotherapy to elderly or nutritionally vulnerable patients is challenging (12,13). If simple, objective preoperative indices could identify patients who are unlikely to benefit from platinum-based regimens, clinicians could avoid exposing these patients to treatment-related toxicities and instead prioritize alternative or modified perioperative strategies, such as EGFR-TKIs or ICIs in appropriate candidates. Evidence directly addressing this question in the setting of resected stage II–IIIA NSCLC is limited.

Objective
The objective of this study was to clarify whether the efficacy of platinum-based adjuvant chemotherapy differs between undernourished and non-undernourished patients with resectable pathological stage IIA–IIIA NSCLC. We specifically evaluated several preoperative nutritional and inflammatory indices, with particular focus on the CONUT score, as potential predictors of survival benefit from platinum-based adjuvant chemotherapy. We present this article in accordance with the STROBE reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-1-2607/rc).

Methods

Ethical considerations
This retrospective cohort study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments and was approved by the institutional review board of Tohoku University Hospital (Sendai, Japan; approval date: 28 September 2020; No. 2021-1-912-1). The board approved a retrospective review of a prospectively maintained database and waived the requirement for individual informed consent.

Study design and setting
We performed a single-center retrospective cohort study using data from a prospective institutional database of thoracic surgical procedures at Tohoku University Hospital, a tertiary care academic center in Sendai, Japan. All consecutive patients who met the eligibility criteria during the study period were included. The study population consisted of patients who underwent lung resection between January 2010 and December 2018, and follow-up was continued until May 2022.

Participants
This analysis represents an additional investigation based on a cohort previously presented at the 2023 World Conference on Lung Cancer (14). We identified patients who underwent lung resection for NSCLC during the study period and were diagnosed with pathological stage IIA–IIIA NSCLC according to the eighth edition of the tumor-node-metastasis (TNM) classification (15).
Inclusion criteria were: histologically confirmed NSCLC; complete anatomical lobectomy with radical intent; pathological stage IIA–IIIA disease; and availability of preoperative laboratory data required to calculate nutritional indices. We excluded patients who received oral anticancer drugs as adjuvant therapy (e.g., tegafur-uracil or other oral agents); and patients with incomplete clinicopathologic or follow-up data for OS. The decision to administer platinum-based adjuvant chemotherapy was made at the discretion of the treating physician based on patient age, PS, comorbidities, and pathological risk factors. Patients were categorized into a surgery-only group and a platinum-based adjuvant chemotherapy group for analysis. All consecutive patients who fulfilled the eligibility criteria were included; no formal sample size calculation was performed.

Follow-up and outcome definition
Patients were routinely followed every three months for the first two postoperative years and typically every six months thereafter. Follow-up evaluations included physical examination, chest radiography, blood tests (including tumor markers), and chest computed tomography (CT). Brain magnetic resonance imaging (MRI) and positron emission tomography/CT (PET/CT) were performed when clinically indicated by symptoms or suspicious findings.
The primary outcome was OS, defined as the interval from the date of surgery to death from any cause or censoring at the date of last follow-up. The follow-up period ended in May 2022. Patients who were alive without a confirmed vital status at the end of follow-up were censored at the date of last contact. Recurrence was diagnosed based on compatible clinical and imaging findings and confirmed histologically when feasible, but recurrence patterns were not the primary endpoint of this analysis.

Clinicopathologic and nutritional assessments
Clinicopathologic data were obtained from electronic medical records and institutional databases. Clinical and pathological staging were re-evaluated according to the eighth edition of the TNM classification. Histological subtypes were classified according to the World Health Organization classification of lung tumors (16).
Comorbidities at the time of surgery were assessed using the Charlson comorbidity index (CCI) (17) and categorized as 0, 1, or ≥2. Smoking history was recorded using the smoking index (<600 vs. ≥600). Pulmonary function was evaluated by the ratio of forced expiratory volume in one second to forced vital capacity (FEV1/FVC), categorized as ≥70% vs. <70%.
Preoperative nutritional and inflammatory status was assessed using blood samples obtained within four days before surgery, based on the following indices, with cutoff values based on previous reports: CAR (18), NLR (19), PLR (20), PNI (21), and CONUT score (9). The CONUT score is calculated from serum albumin, total cholesterol, and total lymphocyte count, and categorizes patients as normal [0–1], mildly [2–4], moderately [5–8], or severely [9–12] undernourished. In this study, a CONUT score of 0–1 was defined as low (normal nutritional status) and ≥2 as high (poor nutritional status), consistent with previous literature. For each nutritional index, patients were classified into normal and undernourished groups according to established cutoffs.

Adjuvant chemotherapy
Platinum-based adjuvant chemotherapy regimens were selected by attending physicians according to contemporaneous guidelines and patient tolerance. In the present cohort, all patients received a cisplatin plus vinorelbine regimen, consisting of cisplatin administered at a dose of 80 mg/m2 on day 1 and vinorelbine administered at a dose of 25 mg/m2 on days 1 and 8 of each cycle. Completion of adjuvant chemotherapy was defined as the administration of four planned cycles. The number of planned cycles and actual completion of therapy were recorded. The completion rate of adjuvant therapy was compared between groups stratified by CONUT score to assess whether nutritional status influenced treatment feasibility.

Statistical analysis
Patient characteristics were summarized using frequencies and proportions for categorical variables and medians with ranges for continuous variables. Baseline characteristics were compared between low and high CONUT groups using the chi-squared test for categorical variables and the Wilcoxon rank-sum test for continuous variables.
OS was estimated using the Kaplan-Meier method, and survival curves were compared using the log-rank test. To evaluate the effect of platinum-based adjuvant chemotherapy within strata of nutritional status, we calculated hazard ratios (HRs) and 95% confidence intervals (CIs) using Cox proportional hazards models. Univariable and multivariable models were constructed with the following covariates, selected a priori as potential prognostic factors: age (<75 vs. ≥75 years), sex (male vs. female), PS (0 vs. 1–3), smoking index (<600 vs. ≥600), FEV1/FVC (≥70% vs. <70%), CCI (0, 1, or ≥2), histology (adenocarcinoma vs. others), pathological stage (II vs. IIIA), platinum-based adjuvant therapy (absent vs. present), pleural invasion (Pl0 vs. Pl1–3), lymphatic invasion (Ly0 vs. Ly1), vascular invasion (V0 vs. V1–2), and pulmonary metastasis (pm0 vs. pm1–2). Proportional hazards assumptions were assessed using log-log survival plots. Patients without a confirmed vital status at the end of follow-up were censored at the date of last contact. All statistical analyses were performed using JMP Pro version 16.2.0 (SAS Institute, Cary, NC, USA). Statistical significance was defined as a two-sided P value <0.05 without adjustment for multiple testing.

Results

Patient selection and baseline characteristics
Among 776 patients who underwent lung resection for NSCLC between January 2010 and December 2018, 154 patients with pathological stage IIA–IIIA NSCLC who underwent radical lobectomy were identified. After excluding patients who received oral anticancer treatment, 135 patients were included in the present analysis. A flow diagram of patient selection is presented in Figure 1. There were no missing data for key clinicopathologic variables and nutritional indices.
Baseline clinicopathologic characteristics are summarized in Table 1. The median age was 70 years (range, 40–86 years), and 92 (68.1%) patients were male. Fourteen patients had a PS of ≥1, 82 had a smoking index of ≥600, 41 had FEV1/FVC <70%, and 36 had a CCI ≥2. Adenocarcinoma was the most common histologic subtype (83 patients, 61.5%), with 52 (38.5%) patients having other histologic types. Pathological stage was II in 75 patients and IIIA in 60.
Platinum-based adjuvant chemotherapy was administered to 68 (50.4%) patients at the discretion of the treating physicians, and the overall completion rate of adjuvant therapy was 63.2%. Among patients who received adjuvant chemotherapy, there were no significant differences between the low and high CONUT groups in the number of administered cycles or in the completion rate of the planned four cycles (Table S1). With respect to nutritional status, 64 (47.4%) patients were classified as undernourished by CAR, 84 (62.2%) by NLR, 52 (38.5%) by PLR, 84 (62.2%) by PNI, and 62 (45.9%) by CONUT score. When stratified by CONUT, the high CONUT group had a lower proportion of adenocarcinomas than the low CONUT group, whereas other baseline characteristics, including the completion rate of adjuvant therapy, did not differ substantially between groups.

Survival according to CONUT score and adjuvant chemotherapy
The median follow-up durations were 4.1 years for all patients and 4.9 years for censored patients. Figure 2 shows OS according to adjuvant chemotherapy status within CONUT-defined nutritional strata.
Among patients with low CONUT scores (normal nutritional status), the adjuvant chemotherapy group had significantly better OS than the surgery-only group (HR 0.318, 95% CI: 0.137–0.737; P=0.008). In contrast, among patients categorized as undernourished by CONUT (high CONUT group), OS did not differ significantly between the surgery-only and adjuvant chemotherapy groups (HR 0.539, 95% CI: 0.227–1.277; P=0.16).
When other nutritional indices (CAR, NLR, PLR, PNI) were used to stratify patients, OS tended to be better in the adjuvant chemotherapy group than in the surgery-only group, regardless of nutritional status, but these differences were not statistically significant (Table 2). Overall, the CONUT score most clearly discriminated between patients who derived benefit from adjuvant chemotherapy and those who did not. Among patients who completed platinum-based adjuvant chemotherapy, recurrence rates were comparable between the low and high CONUT groups; however, both all-cause mortality and lung cancer-related mortality were higher in the high CONUT group (Table S2).

Multivariable analysis
Multivariable Cox regression analyses stratified by CONUT score are summarized in Table 3 and Table S3. In the low CONUT group, platinum-based adjuvant therapy remained an independent prognostic factor for OS (HR 0.065, 95% CI: 0.010–0.423; P=0.004) after adjustment for clinicopathologic variables. In contrast, in the high CONUT group, platinum-based adjuvant therapy was not associated with improved OS (HR 0.508, 95% CI: 0.200–1.291; P=0.15), and no independent prognostic factors were identified in either univariable or multivariable analyses.
Figure 3 shows forest plots of HRs for adjuvant chemotherapy in the overall cohort and in the low and high CONUT subgroups. Although the P value for the interaction between CONUT category and the effect of adjuvant chemotherapy was not statistically significant, there was a clear trend toward reduced benefit of adjuvant chemotherapy in the high CONUT group.

Discussion

Key findings
In this single-center retrospective cohort study of patients with resectable pathological stage IIA–IIIA NSCLC, we evaluated whether preoperative nutritional status modifies the survival benefit of platinum-based adjuvant chemotherapy. Among several nutritional indices examined, the CONUT score most effectively discriminated the benefit of adjuvant chemotherapy. Patients with low CONUT scores experienced a significant survival advantage from platinum-based adjuvant chemotherapy, whereas those with high CONUT scores did not.

Strengths and limitations
The strengths of this study include the use of a prospectively maintained surgical database, clearly defined inclusion and exclusion criteria, and systematic evaluation of multiple established nutritional and inflammatory indices. We also applied multivariable Cox models with adjustment for a comprehensive set of prognostic factors, including age, PS, comorbidities, pulmonary function, histology, and pathological stage.
However, several limitations should be acknowledged. First, the retrospective, single-center design introduces the potential for selection bias and residual confounding. Although multivariable analyses suggested limited selection bias in the allocation of adjuvant chemotherapy, treatment decisions were made by individual physicians, and unmeasured confounders may remain. Second, the sample size—particularly in subgroup analyses stratified by nutritional indices—was relatively small, which may have limited statistical power. The lack of a statistically significant interaction between CONUT category and treatment effect may partly reflect this limitation. Third, OS was selected as the primary endpoint to capture the net clinical benefit of adjuvant chemotherapy, balancing potential oncologic efficacy against treatment-related toxicity and host-related vulnerability. Although recurrence patterns were comparable between CONUT groups, patients with high CONUT scores experienced a higher risk of lung cancer–related death, suggesting impaired resilience after recurrence or during subsequent treatments. Nevertheless, recurrence-free survival and competing-risk analyses were not evaluated in detail, which should be addressed in future prospective studies. Finally, we did not assess dynamic changes in nutritional status over time, which might also influence treatment tolerance and outcomes.

Comparison with similar research
Previous meta-analyses and pooled analyses have established the survival benefit of platinum-based adjuvant chemotherapy in resected NSCLC among fit patients (3,4). Nutritional indices including CONUT have been reported as prognostic markers in various malignancies, such as urothelial carcinoma, esophageal cancer, hepatocellular carcinoma, gastric cancer, and lung cancer (22-27). Our findings are consistent with prior reports that poor nutritional status is associated with adverse outcomes; however, this study extends the literature by suggesting that CONUT may not only be prognostic but also predictive of the magnitude of benefit from platinum-based adjuvant chemotherapy in resected stage II–IIIA NSCLC.

Explanations of findings
The CONUT score integrates serum albumin, total cholesterol, and total lymphocyte count, reflecting both nutritional reserves and systemic immune-inflammatory status. Unlike other inflammatory or nutritional indices such as NLR, PLR, CAR, or PNI, the CONUT score incorporates serum cholesterol in addition to albumin and lymphocyte count, enabling a more comprehensive assessment of host nutritional, immunological, and metabolic reserves. Serum cholesterol is a fundamental component of cellular membranes and lipid rafts, which are essential for immune cell activation and signal transduction. It is also involved in steroid hormone synthesis and energy homeostasis, factors that may influence both tolerance to cytotoxic chemotherapy and recovery from treatment-related toxicity.
In contrast, indices such as NLR and PLR primarily reflect systemic inflammation at a single time point and may be affected by transient inflammatory conditions, potentially limiting their ability to capture long-term physiological resilience. PNI, while reflecting nutritional and immune status, does not account for lipid metabolism. These biological differences may partly explain why the CONUT score more effectively discriminated the survival benefit of platinum-based adjuvant chemotherapy in our cohort.
Patients with high CONUT scores may have diminished physiological reserves, impaired immune function, and altered drug metabolism, which could attenuate the therapeutic effect of cytotoxic chemotherapy or increase susceptibility to treatment-related toxicity. In our cohort, the completion rate of adjuvant chemotherapy and conventional prognostic factors such as PS and CCI were similar between low and high CONUT groups, suggesting that the limited benefit observed in high CONUT patients is not fully explained by overt selection bias or overt frailty. Despite comparable recurrence rates between CONUT groups, a higher risk of lung cancer-related mortality was observed in patients with high CONUT scores, which may reflect impaired resilience following recurrence or reduced tolerance to subsequent therapies. Instead, underlying malnutrition and immune dysfunction captured by the CONUT score may have contributed to reduced net survival benefit from platinum-based chemotherapy in these patients.

Implications and actions needed
The therapeutic landscape of perioperative systemic therapy in NSCLC is rapidly evolving, with EGFR-TKIs and ICIs being incorporated into treatment algorithms for selected molecularly or immunologically defined subgroups. Our results suggest that the CONUT score, a simple preoperative index derived from routine laboratory tests, may help individualize perioperative treatment selection. For patients with low CONUT scores, platinum-based adjuvant chemotherapy should be actively considered because they appear to derive substantial survival benefit. For patients with high CONUT scores, clinicians should carefully weigh the limited expected benefit against potential toxicities and may consider alternative or modified perioperative strategies, including EGFR-TKIs or ICIs in appropriate candidates, or intensive nutritional and supportive interventions before systemic therapy. However, some reports suggest that malnutrition and systemic inflammation may be associated with inferior outcomes and potentially reduced efficacy of ICIs; notably, Shiroyama et al. reported poorer clinical outcomes in patients with sarcopenia treated with PD-1 inhibitors, highlighting the impact of compromised nutritional and physiological status on immunotherapy efficacy (28). Therefore, ICIs should not be assumed to uniformly overcome the biological disadvantages captured by a high CONUT score, and their use in this population should be carefully individualized.
Prospective, multicenter studies are warranted to validate the predictive value of the CONUT score and to determine whether interventions aimed at improving nutritional status can enhance the efficacy and tolerability of perioperative systemic therapies.

Conclusions
The preoperative CONUT score appears to be a useful tool for selecting patients with resectable pathological stage IIA–IIIA NSCLC who are likely to benefit from platinum-based adjuvant chemotherapy. Patients with low CONUT scores derive substantial survival benefit from adjuvant chemotherapy, whereas those with high CONUT scores may derive less survival benefit. Incorporating CONUT-based risk stratification into perioperative decision-making may help optimize systemic treatment strategies in the era of ICIs and molecular targeted agents, particularly for elderly or nutritionally vulnerable patients.

Supplementary
The article’s supplementary files as