Between April 18, 2022, and June 30, 2023, a total of 151 patients with CUP were registered from 66 medical institutions (Fig. 1). All registered patients met the eligibility criteria and were included in the SAF. Seven patients were excluded from the EAS: one was diagnosed with another carcinoma after nivolumab administration, two received off-label short-term nivolumab treatment, and four received concomitant antineoplastic drugs. The EAS, therefore, comprised 144 patients.

Patient disposition. CRF case report form, EAS effectiveness analysis set, SAF safety analysis set

In the SAF (n = 151), the median age was 70 years (range, 26–89 years) and 54 patients (35.8%) were aged ≥ 75 years (Table 1). ECOG PS was 0–1 in 120 patients (79.5%). The most common histologic type was poorly differentiated adenocarcinoma/undifferentiated carcinoma (n = 47; 31.1%), followed by well/moderately differentiated adenocarcinoma (n = 39; 25.8%), squamous cell carcinoma (n = 23; 15.2%), poorly differentiated malignant neoplasm (n = 18; 11.9%), neuroendocrine tumor (n = 2; 1.3%), and other histologic types (n = 22; 14.6%). At the start of nivolumab treatment, 41 patients (27.2%) had lesions confined to lymph nodes only. Seventy-three patients (48.3%) had not received any prior systemic therapy.

The median number of nivolumab administrations was 5.0 (range 1–14) (Supplementary Table S1). The median treatment duration was 73.0 days (range 1–183). At 6 months, 40 patients (26.5%) were still receiving nivolumab, while 111 patients (73.5%) had discontinued treatment. The most common reason for discontinuation was disease progression, including death, in 64 patients (57.7%). Other reasons included lack of expected efficacy (26.1%), occurrence of AEs (19.8%), transfer to another facility (6.3%), and other reasons (5.4%).

In the SAF, TRAEs of any grade occurred in 38 patients (25.2%) and Grade ≥ 3 TRAEs were reported in 18 patients (11.9%) (Table 2). At the System Organ Class level, the incidences of hepatobiliary disorders and immune system disorders were numerically higher in this PMS (3.3% and 0.7%, respectively) than in the NM-K2002 study (1.8% and 0%, respectively; Fig. 2). At the Preferred Term, events occurring in ≥ 2.5% of patients included interstitial lung disease (ILD) (3.3%), hepatic function abnormal (2.6%), and hypothyroidism (2.6%). The most common Grade ≥ 3 TRAEs were hepatic function abnormal (2.0%), ILD (1.3%), and stomatitis (1.3%). One treatment-related death occurred due to neutropenia.

Comparison of TRAE incidence between this PMS and NM-K2002. PMS post-marketing surveillance, TRAE treatment-related adverse event

In subgroup analyses, the incidences of TRAEs were 25.0% in patients with an ECOG PS of 0–1 and 25.8% in those with an ECOG PS of 2–4, with rates of Grade ≥ 3 TRAEs of 12.5% and 9.7%, respectively. By age, the incidences of TRAEs were 33.3% in patients aged ≥ 75 years and 20.6% in those aged < 75 years, with Grade ≥ 3 TRAEs occurring in 18.5% and 8.2%, respectively. Among patients aged ≥ 75 years, ILD (7.4%), hepatic function abnormal (3.7%), and hypothyroidism (3.7%) were the most frequently observed TRAEs. The incidences of TRAEs in patients with a history of pulmonary disease were 71.4% for any grade TRAEs and 57.1% for Grade  ≥ 3 TRAEs, compared with 23.1% and 9.8% in those without such a history. In patients with a history of thyroid disease, the incidences were 66.7% for any grade TRAEs and 16.7% for Grade ≥ 3 TRAEs, compared with 23.6% and 11.8% in those without such a history (Supplementary Table S2).

TRAEs of special interest with an incidence of  ≥ 3% included hepatitis-related events (hepatitis fulminant, hepatic failure, hepatic function disorder, hepatitis, and cholangitis sclerosing), which occurred in 5.3% of patients (Grade ≥ 3 TRAEs: 2.6%), ILD in 4.6% (Grade ≥ 3 TRAEs: 1.3%), and endocrine disorders (e.g., thyroid dysfunction, pituitary insufficiency, and adrenal disorder) in 4.6% (Grade ≥ 3 TRAEs: 1.3%) (Table 3). Recovery or improvement was reported in 75.0% of patients with hepatitis-related TRAEs, all patients with ILD, and in 28.6% of patients with endocrine disorders (Supplementary Table S3). In addition, encephalitis and meningitis (0.7%), and type 1 diabetes mellitus (0.7%) were observed in patients with CUP in this PMS. For each of these categories, Grade  ≥ 3 TRAEs occurred in 0.7% of patients. The patient with encephalitis and meningitis recovered with sequelae, and the patient with type 1 diabetes mellitus had not recovered.

Among the 144 patients in the EAS, the best overall responses were CR in 1.4%, PR in 20.1%, SD in 25.0%, PD in 45.1%, and NE in 8.3%. The ORR was 21.5% (95% confidence interval [CI], 15.1–29.1) and the DCR was 46.5% (95% CI, 38.2–55.0).

By ECOG PS, the ORR was 25.4% (95% CI, 17.7–34.4) for PS 0–1, 8.0% (95% CI, 1.0–26.0) for PS 2, and 0% (95% CI, 0.0–52.2) for PS 3–4. By age group, the ORR was 24.7% (95% CI, 16.4–34.8) in patients aged < 75 years and 15.7% (95% CI, 7.0–28.6) in those aged ≥ 75 years. The ORR was 25.4% (95% CI, 15.5–37.5) without prior systemic therapy and 18.2% (95% CI, 10.3–28.6) with prior systemic therapy. By histology, the ORR was 10.8% (95% CI, 3.0–25.4) for well/moderately differentiated adenocarcinoma, 15.6% (95% CI, 6.5–29.5) for poorly differentiated adenocarcinoma/undifferentiated carcinoma, 27.3% (95% CI, 10.7–50.2) for squamous cell carcinoma, 50.0% (95% CI, 1.3–98.7) for neuroendocrine tumor, 41.2% (95% CI, 18.4–67.1) for poorly differentiated malignant neoplasm, and 28.6% (95% CI, 11.3–52.2) for other tumors. The ORR was 36.6% (95% CI, 22.1–53.1) among patients with lymph node-only involvement compared with 15.5% (95% CI, 9.1–24.0) in patients with disease at other sites, with or without lymph node involvement (Fig. 3). The corresponding DCRs in these subgroups are also presented in Supplementary Table S4.

ORR according to baseline patient characteristics in patients with CUP who received nivolumab (n = 144). a “LN-only” denotes lymph node-only involvement, defined as no lesions at sites other than lymph node(s). b “Other sites” denotes lesions at sites other than lymph nodes, with or without lymph node involvement. CI confidence interval, CUP cancer of unknown primary, ECOG PS Eastern Cooperative Oncology Group performance status, LN lymph node, ORR overall response rate

The 6-month OS rate was 71.0% (95% CI, 62.5–78.0) among all patients, and the median OS was not reached (Fig. 4). By ECOG PS, the 6-month OS rate was 81.2% (95% CI, 72.3–87.4) for ECOG PS 0–1 (median not reached), 36.6% (95% CI, 17.0–56.5) for ECOG PS 2 (median 125.0 days [95% CI, 60.0–not reached]), and not reached for ECOG PS 3–4 (median 24.0 days [95% CI, 6.0–not reached]) (Supplementary Figure S1). By age, the 6-month OS rate was 72.6% (95% CI, 61.9–80.7) for patients < 75 years and 68.4% (95% CI, 53.0–79.7) for those aged ≥ 75 years (Supplementary Figure S2). By treatment history, the 6-month OS rate was 72.4% (95% CI, 59.3–81.9) without prior systemic therapy and 69.9% (95% CI, 58.0–79.1) with prior systemic therapy (Supplementary Figure S3). By histology, the 6-month OS rates were 63.2% (95% CI, 45.1–76.8) for well/moderately differentiated adenocarcinoma, 61.9% (95% CI, 44.7–75.2) for poorly differentiated adenocarcinoma/undifferentiated carcinoma, 81.8% (95% CI, 58.5–92.8) for squamous cell carcinoma, 100.0% (95% CI, 100.0–100.0) for neuroendocrine tumor, 81.1% (95% CI, 51.9–93.5) for poorly differentiated malignant neoplasm, and 79.9% (95% CI, 54.8–92.0) for other tumors (Supplementary Figure S4). Patients with lymph node-only involvement had a 6-month OS rate of 92.4% (95% CI, 78.3–97.5) compared with 62.0% (95% CI, 51.3–71.0) in patients with disease at other sites, with or without lymph node involvement (Supplementary Figure S5).

Kaplan–Meier plot of OS in patients with CUP who received nivolumab (n = 144). CI confidence interval, CUP cancer of unknown primary, NR not reached, OS overall survival