This multicenter retrospective observational study investigated patients who underwent NIR-PIT for head and neck cancer. Okayama University Hospital served as the coordinating center, and the Department of Otolaryngology–Head and Neck Surgery at Kumamoto University Hospital and the Department of Otolaryngology at Japanese Red Cross Society Himeji Hospital participated in this study.

Eligible patients were those who underwent initial NIR-PIT for recurrent or metastatic HNC between January 2021 and March 2026. Data regarding patient characteristics, tumor-related factors, treatment details, clinical course, and outcomes were retrospectively collected from the medical records of each institution. Follow-up data were collected through June 2026.

For the primary analysis, patients were classified according to prior exposure to ICIs before NIR-PIT. Prior ICI exposure was defined as any history of ICI administration before the initial NIR-PIT session.

In an exploratory analysis, NIR-PIT-treated lesions were categorized according to treatment sequence and timing of ICI administration: G1, lesions with prior ICI exposure before NIR-PIT; G2, lesions treated with NIR-PIT alone; G3, lesions exposed to ICIs before disease progression after NIR-PIT; and G4, lesions exposed to ICIs after disease progression following NIR-PIT. The G3 group included lesions from patients who received ICIs as planned treatment or early sequential therapy after NIR-PIT. The G4 group included lesions from patients who started ICI therapy after confirmed progression of the NIR-PIT-targeted lesion. Disease progression for group classification was defined as radiologically confirmed progression of the NIR-PIT-targeted lesion.

The following variables were collected: age, sex, performance status (PS; ECOG), primary tumor site, presence of distant metastasis, date of initial NIR-PIT, number of NIR-PIT sessions, type of ICI, date of ICI initiation, tumor response, occurrence and date of local progression, date of subsequent treatment initiation, date of last follow-up, and survival status.

Tumor response was assessed based on radiological evaluations performed after NIR-PIT. Computed tomography or magnetic resonance imaging was used for imaging assessment, and treatment response was evaluated according to the Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1 [14]. Radiological assessments, including both RECIST response evaluation and assessment of local progression, were performed by the treating investigators at each participating institution based on routine clinical imaging. No independent central radiological review was performed.

Tumor response was categorized as complete response (CR), partial response (PR), stable disease (SD), or progressive disease (PD). Best overall response (BOR) was defined as the best response recorded during the observation period. Objective response rate (ORR) was defined as the proportion of lesions achieving CR or PR, whereas disease control rate (DCR) was defined as the proportion achieving CR, PR, or SD.

ORR, DCR, and local control were evaluated on a per-lesion basis, whereas overall survival (OS) was evaluated on a per-patient basis.

The primary endpoint was local control, which was defined as the interval between the date of initial NIR-PIT and the date of confirmed local progression. Local progression was defined as radiological enlargement or regrowth of the NIR-PIT-targeted lesion. Unlike RECIST-defined progressive disease, which reflects overall disease status, local progression was evaluated specifically for the NIR-PIT-targeted lesion because local control was the primary endpoint of this study. Consequently, progression at non-target lesions or distant metastatic sites according to RECIST was not considered a local progression event unless enlargement or regrowth of the NIR-PIT-targeted lesion was confirmed. Recurrence of the same lesion after achievement of CR was also considered local progression. Lesions without local progression were censored at the earlier of either the date of subsequent treatment initiation or the date of last follow-up. Subsequent treatment initiation was selected as a censoring event because additional therapies could directly influence local tumor status and confound assessment of local control attributable to NIR-PIT. Death was treated as a censoring event rather than a competing event. Secondary endpoints were ORR and DCR.

OS was analyzed as an exploratory endpoint. OS was defined as the interval between the date of initial NIR-PIT and the date of death from any cause. Patients who were alive as of the last follow-up were censored.

Categorical variables were compared using the chi-square test or Fisher’s exact test, as appropriate. The same methods were used to compare local progression rate, ORR, and DCR between groups.

Local control and OS were estimated using Kaplan–Meier methods and compared using the log-rank test. Univariate Cox proportional hazards analysis was performed to explore the association between prior ICI exposure and local control.

All statistical tests were two-sided, with values of P < 0.05 considered significant. Because this was an exploratory retrospective study with a limited sample size and no clear candidate variables for multivariable analysis were evident from the univariate analyses, multivariable analysis was not performed. Median duration of follow-up was estimated using the reverse Kaplan–Meier method. All statistical analyses were performed using IBM SPSS Statistics for Windows, version 26.0 (IBM Corp., Armonk, NY, USA).

All study protocols were approved by the institutional review board at Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Sciences, Japan (approval no. K2605-019). The study was conducted in accordance with the Declaration of Helsinki and institutional data-protection policies.