Oxaliplatin-containing regimens are widely used as standard first-line chemotherapy for advanced gastric cancer and are recommended in international clinical guidelines (1). However, chemotherapy-induced peripheral neuropathy (CIPN) is a major dose-limiting toxicity that develops and worsens in a dose-dependent manner, often requiring dose reduction or treatment discontinuation (2, 3).

Oxaliplatin is a major cause of CIPN, and the risk of its development and progression increases with cumulative exposure (2, 4). Oxaliplatin-induced neuropathy can adversely affect quality of life, treatment continuity, the selection of subsequent chemotherapy, and may persist long after treatment completion (3, 5, 6). Paclitaxel, which is widely used as second-line therapy, is also associated with clinically significant neurotoxicity, making sequential treatment particularly challenging (7, 8).

Patients receiving paclitaxel after oxaliplatin-based chemotherapy often present with varying degrees of residual neuropathy at treatment initiation. Previous studies have shown that patients with pre-existing CIPN are more likely to develop severe neuropathy during paclitaxel treatment (9, 10). However, whether the risk of subsequent neuropathy is primarily determined by cumulative oxaliplatin exposure or by the patient’s neurological status at paclitaxel initiation remains unclear.

Although cumulative oxaliplatin exposure is widely recognized as an important determinant of neurotoxicity, previous studies have largely relied on dichotomous assessments of baseline CIPN, and the relative contributions of cumulative oxaliplatin exposure and baseline neurological status have not been adequately evaluated (9, 10). Clarifying which factor better predicts clinically significant neuropathy may facilitate risk stratification and individualized management during sequential chemotherapy.

Therefore, this study aimed to identify factors associated with clinically significant CIPN during paclitaxel treatment, with particular emphasis on determining whether baseline CIPN or cumulative oxaliplatin exposure more accurately predicts subsequent neurotoxicity.

Study population. This single-center retrospective observational study included 76 patients with advanced gastric cancer who received oxaliplatin-containing chemotherapy as first-line treatment followed by paclitaxel-containing therapy as second-line treatment at our Institution. The study period was from January 2021 to October 2025. No formal sample size calculation was performed because this was a retrospective observational study, and all eligible patients treated during the study period were included. Patients were eligible if they had advanced gastric cancer, received oxaliplatin-containing chemotherapy as first-line treatment, and subsequently received paclitaxel-containing therapy as second-line treatment at our Institution between January 2021 and October 2025.

Data collection. The following data were extracted from medical records: age, sex, serum albumin level, history of diabetes mellitus, renal function (creatinine clearance), cumulative oxaliplatin dose during first-line therapy, CIPN grade at paclitaxel initiation, paclitaxel treatment regimen, clinical course of CIPN (worsening, maximum grade, and development of grade ≥2 CIPN), and dose reduction or treatment discontinuation due to CIPN. Baseline variables, including age, serum albumin level, creatinine clearance, diabetes mellitus status, and CIPN grade, were assessed at paclitaxel initiation. Cumulative oxaliplatin dose was calculated over the course of first-line treatment.

Endpoints. CIPN was evaluated according to the Common Terminology Criteria for Adverse Events version 5.0 (11). The primary endpoint was the development of grade ≥2 CIPN. Secondary endpoints included CIPN worsening (defined as an increase of at least one grade), dose reduction due to CIPN, and treatment discontinuation due to CIPN.

Time to grade ≥2 CIPN and time to CIPN worsening were defined as the number of days from paclitaxel initiation to each event. Patients without adverse events were censored at the last follow-up. Patients with grade ≥2 CIPN at paclitaxel initiation were excluded from the analysis of grade ≥2 CIPN development.

Statistical analysis. Continuous variables are presented as median [interquartile range (IQR)] and categorical variables as number with percentage. For univariate analyses, the Mann–Whitney U-test was used for continuous variables and Fisher’s exact test for categorical variables.

Time-to-event analyses were performed using the Kaplan–Meier method, and differences between groups were evaluated using the log-rank test. Patients were stratified by the presence or absence of baseline CIPN at paclitaxel initiation and according to the median cumulative oxaliplatin dose.

Multivariate analyses were conducted using Cox proportional hazards models. The following variables were included as covariates: cumulative oxaliplatin dose (per 100 mg/m² increment), CIPN at paclitaxel initiation (grade 1 vs. 0), age, serum albumin level, and history of diabetes mellitus.

Logistic regression analyses were performed to identify factors associated with the development of grade ≥2 CIPN and dose reduction due to CIPN. Logistic regression analysis for CIPN worsening was not performed because of concerns regarding the model stability.

Statistical analyses were performed using EZR version 1.60 (Saitama Medical Center, Jichi Medical University) (12).

Ethical considerations. This study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Ethics Committee (approval number: 20251127-21). As this was a retrospective observational study, informed consent was obtained using an opt-out approach.

Patient characteristics. Patient characteristics are summarized in Table I. The median age was 73 (IQR=66-79) years, and the cohort included 54 men and 22 women. The median serum albumin level was 3.5 (IQR=3.2-3.8) g/dl, and the median creatinine clearance was 61.6 (IQR=52.4-82.8) ml/min. Nineteen patients (25.0%) had a history of diabetes mellitus. Paclitaxel was administered in combination with ramucirumab in 66 (86.8%) patients. The median cumulative oxaliplatin dose during first-line therapy was 1011 (IQR=697.3-1,362) mg/m². At paclitaxel initiation, baseline CIPN was grade 0 in 32 patients (42.1%), grade 1 in 42 (55.3%), and grade ≥2 in 2 (2.6%).

Primary endpoint: development of grade ≥2 CIPN. The results of the univariate analysis for factors associated with the development of grade ≥2 CIPN are presented in Table II. The proportion of patients receiving paclitaxel plus ramucirumab combination therapy was lower in the group with grade ≥2 CIPN group (76.5% vs. 95.2%, p=0.036). The proportion of patients with CIPN grade ≥1 at paclitaxel initiation was higher in the group with grade ≥2 CIPN, although the difference was not statistically significant (70.6% vs. 47.6%, p=0.062). No other variable showed a significant association.

The results of the Cox proportional hazards analysis for CIPN-related outcomes are presented in Table III. In the multivariate Cox analysis, cumulative oxaliplatin dose was not significantly associated with the development of grade ≥2 CIPN [hazard ratio (HR)=0.96, p=0.350]. In contrast, grade 1 CIPN at paclitaxel initiation was an independent predictor of development of grade ≥2 CIPN (HR=4.23, 95% confidence interval=1.67-10.72; p=0.002). No other variable showed a significant association.

The results of the multivariable logistic regression analysis for CIPN-related outcomes are presented in Table IV. CIPN at paclitaxel initiation was significantly associated with the development of grade ≥2 CIPN [odds ratio (OR)=5.00, 95% confidence interval=1.58-15.80; p=0.006]. The cumulative oxaliplatin dose showed a tendency toward reduced risk, but the trend was not statistically significant (OR=0.91, p=0.065).

Secondary endpoint: worsening of CIPN. When comparing patients by baseline CIPN status, Kaplan–Meier analysis showed that patients with CIPN at paclitaxel initiation tended to have a longer time to CIPN worsening than those without CIPN at baseline; however, the difference was not statistically significant (114 vs. 57 days, p=0.085) (Figure 1). Nor was any significant difference observed according to cumulative oxaliplatin dose (p=0.384) (Figure 2).

None of the evaluated variables showed a significant association with the worsening of CIPN in the univariate analysis.

In the multivariate analysis, cumulative oxaliplatin dose (per 100 mg/m² increase) was not significantly associated with CIPN worsening (HR=1.01, p=0.847). CIPN at paclitaxel initiation and serum albumin level showed tendencies toward reducing the risk of CIPN worsening, although these associations were not statistically significant.

Clinical outcomes: treatment modification due to CIPN. In the logistic regression analysis, cumulative oxaliplatin dose was not associated with dose reduction due to CIPN (OR=0.95, p=0.373). CIPN at paclitaxel initiation tended to increase the risk of dose reduction, although the association was not statistically significant (OR=3.71, p=0.087).

In this study, CIPN at paclitaxel initiation was identified as the strongest predictor of clinically significant neuropathy during second-line paclitaxel treatment following first-line oxaliplatin-based chemotherapy. Patients with grade 1 baseline CIPN had a significantly higher risk of developing grade ≥2 CIPN, whereas cumulative oxaliplatin exposure was not independently associated with neurotoxicity. These findings suggest that neurological status at paclitaxel initiation provides more clinically relevant information for neurotoxicity risk stratification than prior chemotherapy exposure alone. Therefore, routine assessment of baseline CIPN may facilitate individualized treatment selection, dose modification, and monitoring during paclitaxel therapy.

Oxaliplatin induces CIPN in a cumulative, dose-dependent manner, and cumulative exposure above a certain threshold is a well-established risk factor for neurotoxicity (2, 4). Taxane-induced peripheral neuropathy is also considered a cumulative dose-dependent toxicity (13). Recent experimental work has further characterized the temporal development of paclitaxel-induced neurological dysfunction in vivo (14). Furthermore, previous studies in colorectal cancer survivors have demonstrated persistent sensory neuropathy associated with higher cumulative oxaliplatin exposure (3, 4, 15). Therefore, cumulative exposure has generally been regarded as an important determinant of subsequent neurotoxicity.

In contrast to these observations, cumulative oxaliplatin exposure was not significantly associated with either the development or progression of CIPN during paclitaxel treatment in the present study. One possible explanation is time-dependent bias. Patients who developed neuropathy later were more likely to continue oxaliplatin treatment longer and consequently had higher cumulative oxaliplatin doses, potentially attenuating the apparent association between cumulative exposure and subsequent neuropathy. In addition, treatment modifications in routine clinical practice, including dose reduction and interval adjustment, may have influenced neuropathy outcomes. Therefore, our findings should not be interpreted as indicating that cumulative oxaliplatin exposure is unimportant but rather that cumulative exposure alone may not adequately predict neurotoxicity during subsequent paclitaxel therapy.

CIPN before paclitaxel initiation was independently associated with the development of grade ≥2 CIPN in both the Cox proportional hazards and multivariate logistic regression analyses. These findings are consistent with previous reports showing that pre-existing neuropathy increases susceptibility to subsequent taxane-related neurotoxicity (9, 10). More recently, a prospective multicenter study in patients with advanced gastric cancer demonstrated that residual platinum-induced peripheral sensory neuropathy was associated with an increased risk of clinically relevant neurotoxicity during subsequent paclitaxel-based therapy, particularly among patients with more severe neuropathy at baseline (16). Together, these findings support the clinical importance of neurological status at paclitaxel initiation for neurotoxicity risk stratification. Our results further suggest that current neurological status may provide more clinically relevant information than prior cumulative oxaliplatin exposure alone. Because baseline neurological assessment directly reflects the cumulative effects of previous neurotoxic exposure as well as individual susceptibility, it may represent a practical tool for identifying patients at high risk for clinically significant neuropathy during paclitaxel treatment.

Although all patients in this study had an Eastern Cooperative Oncology Group performance status of 0-1, baseline CIPN remained a consistent predictor of neurotoxicity. This finding suggests that neurological assessment provides clinically useful information even in patients with preserved general condition. In addition, higher serum albumin levels tended to be associated with a lower risk of CIPN progression, suggesting that patient-related factors such as nutritional status and physiological reserve may also influence neurotoxicity. These observations are consistent with previous studies demonstrating the importance of host-related factors in predicting treatment toxicity (13, 15, 17).

Study limitations. Firstly, this was a single-center retrospective study with a relatively small sample size, and residual confounding and selection bias cannot be excluded. Secondly, CIPN was evaluated retrospectively using medical records, which may have resulted in variability or underestimation of toxicity grading. A recent review has highlighted the heterogeneity of CIPN assessment methods and the potential value of integrating clinician-rated, patient-reported, and objective measures for more comprehensive evaluation (18). Thirdly, cumulative oxaliplatin exposure was analyzed as a fixed variable rather than as a time-dependent covariate, and therefore time-dependent bias may have influenced the observed associations. Prospective studies with larger sample sizes and time-dependent analytical approaches are warranted to validate these findings and further clarify the relationship between cumulative neurotoxic exposure, baseline neurological status, and subsequent CIPN during sequential chemotherapy.

Baseline CIPN was a stronger predictor of clinically significant neuropathy during paclitaxel treatment than cumulative oxaliplatin exposure. Routine assessment of baseline neurological status before paclitaxel initiation may help identify patients at high risk for neurotoxicity and support individualized treatment selection, dose modification, and monitoring strategies.

The Authors declare no conflicts of interest.

AM and MK contributed to the study design, data collection, and manuscript preparation. AM is the guarantor of the study and all data. YI and EU contributed to the study design, critically revised the manuscript, and supervised the study. All Authors read and approved the final manuscript.

During the preparation of this manuscript, the Authors used ChatGPT (OpenAI, San Francisco, CA, USA) to improve the clarity and readability. The Authors reviewed and edited all AI-assisted content and take full responsibility for the accuracy and integrity of the manuscript.