AUTHORS
Seda S. Tolu, MD, William Britton, MD2, Amarachukwu Nwaoha3, David DeStephano4, Alexander D. Sanjurjo, B.S.5, Yoni Sacknovitz, B.A.6, Sarah Sadoun A Alnafisee, MD MPH MS7, Hua-Jay Jeffery Cherng, MD8, Anna Sarah Erem, MD9, Alejandro Gru, MD10, Jennifer Effie Amengual, MD11, Larisa J. Geskin, MD12, Barbara Pro, MD13
AFFILIATIONS
1 Columbia University Irving Medical Center, New York, United States
2 Columbia University Irving Medical Center, New York, United States
3 Columbia University Irving Medical Center, New York, United States
4 Columbia University Irving Medical Center, New York, United States
5 Columbia University Irving Medical Center, New York, United States
6 Columbia University Irving Medical Center, New York, United States
7 Columbia University Irving Medical Center, New York, United States
8 Columbia University Irving Medical Center, New York, United States
9 Columbia University Irving Medical Center, New York, United States
10 Columbia University Irving Medical Center, New York, United States
11 Columbia University Irving Medical Center, New York, United States
12 Columbia University Irving Medical Center, New York, United States
13 Columbia University Irving Medical Center, New York, United States
ABSTRACT
Introduction: Cutaneous T-cell lymphoma (CTCL) is a chronically relapsing cancer with poor outcomes in advanced stages. Programmed death-1 (PD-1) is variably expressed in CTCL, yet its biologic and prognostic significance remains uncertain. Although PD-1 blockade produces modest responses in relapsed/refractory CTCL, PD-1 expression has not predicted treatment response (Khodadoust 2020). We therefore evaluated whether baseline tumor PD-1 expression identifies clinically distinct CTCL subsets and is associated with disease characteristics and long-term outcomes.
Methods: We conducted a single-center retrospective study of patients with CTCL (mycosis fungoides [MF] or Sézary syndrome [SS]) treated at Columbia University Irving Medical Center between 2010 and 2025 who had an available baseline skin biopsy evaluated for PD-1 expression by immunohistochemistry. PD-1-high was defined as ≥50%, representing expression in the majority of malignant lymphocytes, and PD-1-low as <50%. Additionally, sensitivity analyses modeled PD-1 continuously per 10%-point increase. Primary endpoint was overall survival (OS). Secondary endpoints were time from diagnosis to first systemic treatment (TTFT), and time from first to next systemic treatment (TTNT). Multivariable cox models adjusted for age, stage at diagnosis, large-cell transformation (LCT) at diagnosis, and baseline modified Severity-Weighted Assessment Tool (mSWAT).
Results: Among 179 patients with CTCL and available PD-1 staining, 38 (21.2%) had PD-1-high disease and 141 (78.8%) had PD-1-low disease. Median age was 60.5 versus 58.0 years, and 55.3% versus 56.7% were male, respectively. Median follow-up was 49.5 months (m). Stage was available for 175 patients: 69.7% (n=122) had early stage (IA-IIA) and 30.3% (n=53) had advanced stage (IIB-IV). Compared with the PD-1-low group, patients with PD-1-high disease more frequently had SS (39.5% vs 5.7%; p<0.001) and LCT at diagnosis (14.3% vs 3.0%; p=0.021).
PD-1-high disease was associated with markers of greater baseline disease burden. Advanced-stage disease was present in 64.9% patients with PD-1-high disease compared with 21.0% with PD-1-low disease (p<0.001). Median baseline mSWAT was 45.0 (IQR, 10.0–107.8) in the PD-1-high group versus 8.0 (IQR, 2.0–40.0) in the PD-1-low group (p<0.001). Receipt of first-line systemic therapy (excluding skin-directed alone) was more frequent among PD-1-high disease (76.3%, n=29) than PD-1-low (34%, n=48; p<0.001). Patients with PD-1-high disease had shorter TTFT of 3.0m compared to 91.1m in PD-1-low disease with unadjusted HR 3.17 (95% CI, 2.01–4.99; p<0.001). This association was attenuated after multivariable adjustment ([a]HR 0.92; 95% CI, 0.47–1.79; p=0.810). Median TTNT was numerically shorter in the PD-1-high group versus PD-1-low group (9.5m vs. 14.8m, aHR 1.47; 95% CI, 0.86–2.53; p=0.159). Patients with PD-1-high disease received more systemic lines of treatment (median, 2 vs 0; p<0.001). Notably, 7 patients received treatment with PD-1 blockade (pembrolizumab) including only 2 treated with monotherapy, precluding meaningful analysis of response by PD-1 expression.
PD-1-high disease was associated with inferior OS in the unadjusted analysis (HR 2.50; 95% CI, 1.08–5.82; p=0.033). Median OS was 107.9m in the PD-1-high group and was not reached in the PD-1-low group. This association was no longer significant after multivariable adjustment (aHR, 1.23; 95% CI, 0.40–3.79; p=0.713;). Findings were unchanged in sensitivity models additionally adjusting for CTCL subtype (MF versus SS).
In continuous sensitivity analyses, each 10% increase in expression of PD-1 was associated with worse unadjusted OS (HR, 1.17; 95% CI, 1.03–1.33; p=0.017) and earlier TTFT (HR, 1.18; 95% CI, 1.10–1.27; p<0.001). However, neither association remained significant after adjustment.
Conclusion: High baseline PD-1 expression identifies a clinically distinct subset of CTCL characterized by advanced stage, greater skin disease burden, and earlier need for systemic therapy. Associations between PD-1 expression, TTFT, and OS were attenuated after adjustment for baseline disease severity, suggesting that PD-1 expression may primarily reflect disease burden rather than function as an independent prognostic biomarker. Larger multicenter studies with standardized PD-1 assessment are warranted to clarify its biologic significance and potential role in therapeutic stratification.
DISCLOSURES
Seda S. Tolu, MD: Adcetris - Consultancy (Includes expert testimony); Pfizer - Consultancy (Includes expert testimony)