Abstract
Introduction: Thyroid immune-related adverse events (irAEs) have been reported to have prognostic significance among cancer patients treated with anti-PD1 and anti-PDL1 monotherapies. There are scanty data in the literature thus far about the clinical course and prognostic significance of thyroid irAEs in the routine clinical use of combination anti-PD1/anti-CTLA4 treatment in advanced cancer patients. We evaluated the clinical course and predictors of thyroid irAEs, in relation to outcomes of advanced cancer patients treated with combination anti-PD1/anti-CTLA4. Method: We conducted a territory-wide study and identified advanced cancer patients who received ≥1 cycle of combination anti-PD1/anti-CTLA4 between 2015 and 2019 in Hong Kong. Patients were excluded if (i) they had a history of thyroid disorder or thyroid cancer, (ii) immune checkpoint inhibitor-related endocrinopathies occurred before the commencement of combination anti-PD1/anti-CTLA4, (iii) they were on concurrent tyrosine kinase inhibitor (TKI), (iv) baseline thyroid function tests (TFTs) were absent or abnormal, and (v) the duration of follow-up was <30 days. TFTs were monitored every three weeks. Thyroid irAE was defined by ≥2 abnormal TFTs after initiation of combination anti-PD1/anti-CTLA4 in the absence of other causes. The initial presentation was classified into hypothyroidism (overt if TSH >4.8 mIU/L and fT4 <12 pmol/L; subclinical if TSH >4.8 mIU/L and fT4 12-23 pmol/L) and thyrotoxicosis (overt if TSH <0.35 mIU/L and fT4 >23 pmol/L; subclinical if TSH <0.35 mIU/L and fT4 12-23 pmol/L). Results: One hundred and three patients were included (median age: 59 years; 71.8% men). Around half of patients had hepatocellular carcinoma. About 45% had prior anti-PD1 exposure. Upon median follow-up of 6.8 months, 17 patients (16.5%) developed thyroid irAEs, where 6 initially presented with thyrotoxicosis (overt, n=4; subclinical, n=2), and 11 with hypothyroidism (overt, n=2; subclinical, n=9). Eventually, 10 patients (58.8%) required continuous thyroxine replacement. Systemic steroid was not required in all cases. Prior anti-PD1 exposure (OR 3.67, 95% CI 1.19-11.4, p=0.024) independently predicted thyroid irAEs. Multivariable Cox regression analysis revealed that occurrence of thyroid irAEs was associated with better overall survival (adjusted hazard ratio 0.39, 95% CI 0.19-0.79, p=0.009), independent of prior exposure to anti-PD1 (p=0.386) and prior TKI exposure (p=0.155). Conclusion: Thyroid irAEs are common in routine clinical practice among advanced cancer patients treated with combination anti-PD1/anti-CTLA4, and might have potential prognostic significance. Regular TFT monitoring is advised for timely treatment of thyroid irAEs to prevent potential morbidities.