KEYNOTE-013 and KEYNOTE-170: Improving the Precision of Checkpoint Blockade in Patients With Diffuse Large B-Cell Lymphoma

Get Permission

Primary mediastinal B-cell lymphoma is an uncommon but distinct clinicopathologic variant of diffuse large B-cell lymphoma (DLBCL) that typically presents with an aggressive mediastinal mass and invasion of local structures. Most patients are young, typically in their 20s and 30s, and there is a slight female predilection. Despite an aggressive presentation, substantial gains have occurred in the front-line setting. The use of intensive chemoimmunotherapy regimens, often omitting radiation, leads to high response rates in the vast majority of patients.

Sonali M. Smith, MD, FASCO

Sonali M. Smith, MD, FASCO

Relapsed or Refractory Disease

Unfortunately, the management (and prognosis) of relapsed or refractory disease is much more challenging. The traditional approach of salvage chemotherapy plus high-dose chemotherapy and autologous hematopoietic stem cell rescue (HCT) in chemosensitive patients is effective in some patients, but similar to other aggressive B-cell lymphomas, the role and effectiveness of autologous HCT have waned in the modern era. Compared with other DLBCL subtypes, patients with primary mediastinal B-cell lymphoma often have a worse outcome with HCT,1 and relapsed disease despite autologous HCT remains a significant challenge. Patients not responding to salvage chemoimmunotherapy have a dismal survival of just 15% at 3 years.2 In general, there is initial chemosensitivity in front-line settings, but patients with recurrent disease display significant chemoresistance, and there are limited effective treatment options.

The biology of primary mediastinal B-cell lymphoma is quite unique among DLBCLs and is more akin to classical Hodgkin lymphoma, with strong programmed cell death ligand 1 (PD-L1) expression and consistent CD30 positivity. Similar to classical Hodgkin lymphoma, with primary B-cell lymphoma, there are genomic deregulation and amplification of the 9p24 locus with resultant PD-L1, PD-L2, and JAK overexpression. Targeting CD30 with the antibody-drug conjugate brentuximab vedotin has modest activity, leaving a large unmet need in this field. In a small prospective trial of 15 patients, the overall response rate was less than 15%.3


A recent article by Armand et al4—reviewed in this issue of The ASCO Post—shows the single-agent activity of the programmed cell death protein 1 inhibitor pembrolizumab in relapsed and refractory primary mediastinal B-cell lymphoma in two prospective trials: phase Ib KEYNOTE-013 and phase II KEYNOTE-170. Given the rarity of this disease, the collective 74 patients represent one of the larger series published in the past decade.

Pembrolizumab was delivered at 200 mg every 3 weeks for up to 2 years, and the primary endpoints were safety and overall response rate. It is worth noting that this is a heavily pretreated population of patients with a median of three prior lines of therapy and prior autologous HCT in approximately 75% of patients. As is typical of the disease, the median age is young (31–33 years old), and more than half of patients are female.

The key findings are that 45% to 48% of patients have a radiographic response, with 13% to 33% of patients achieving a complete response, and there are no unexpected safety signals for checkpoint blockade. With the caveat of short follow-up, particularly for the larger KEYNOTE-170 trial, the median duration of response has not yet been reached among responders.

“Perhaps one of the more exciting observations in the study is that six patients in the two trials have completed the 2 years of planned therapy and remain in remission off treatment.”
— Sonali M. Smith, MD, FASCO

Tweet this quote

Perhaps one of the more exciting observations in the study is that six patients in the two trials have completed the 2 years of planned therapy and remain in remission off treatment; furthermore, two patients in KEYNOTE-013 remain in remission more than 1 year off therapy. Despite a highly encouraging duration of response, especially for complete responders, progression-free survival is relatively short at 10.4 months and 5.5 months in KEYNOTE-013 and KEYNOTE-170, respectively. Nonresponders have dismal outcomes, with 6 patients dying within 30 days and 11 patients dying within 60 days of starting the study.

A potential explanation for the differential activity is a link between PD-L1/PD-L2 gene dose, protein expression, and likelihood of response. Gene dose for the PD-L1/PD-L2 loci was categorized as polysomic, copy gain, or amplified. Protein expression was determined via a centrally assessed H-score (product of percent malignant cells positive for PD-L1 and the scored intensity of staining). Not only did the degree of PD-L1/PD-L2 gene deregulation correlate with the H-score, but this was predictive of freedom from disease progression following pembrolizumab monotherapy. This could be an effective predictive tool and needs further validation.

Not all DLBCL subtypes respond to checkpoint blockade, and in this sense, primary mediastinal B-cell lymphoma is unique yet again. Among unselected DLBCL, the single-agent response rate to nivolumab is less than 10%.5 However, there may be a subset of patients with DLBCL harboring PD-L1/PD-L2 gene aberrations that are predictive of response.6

Overall, the studies reported by Armand and colleagues establish the safety profile and single-agent response rate of pembrolizumab in primary mediastinal B-cell lymphoma and support the U.S. Food and Drug Administration approval in 2018. The promising response rate and durability of response in complete responders represent an important advance, and checkpoint blockade may be a platform for future combination studies. In addition, the description of a predictive model for response (via fluorescence in situ hybridization or immunohistochemistry) is exciting and, if validated, will help deliver checkpoint blockade in a more precise manner to patients most likely to benefit.

Dr. Smith is the Elwood V. Jensen Professor in Medicine; Interim Chief, Section of Hematology/Oncology; Director, Lymphoma Program, The University of Chicago Medicine.

DISCLOSURE: Dr. Smith has received honoraria from Genentech and Kite; served as a consultant or advisor for TG Therapeutics, Genentech, Bayer, Seattle Genetics, Nordic Nanovector, Gilead/Kite, Portola, and Pharmacyclics; and has received research funding from Portola, Celgene, Forty-Seven, Curis, TG Therapeutics, Roche, Karyopharm, Novartis, and Acerta.


1. Kuruvilla J, Pintilie M, Tsang R, et al: Salvage chemotherapy and autologous stem cell transplantation are inferior for relapsed or refractory primary mediastinal large B-cell lymphoma compared with diffuse large B-cell lymphoma. Leuk Lymphoma 49:1329-1336, 2008.

2. Vardhana S, Hamlin PA, Yang J, et al: Outcomes of relapsed and refractory primary mediastinal (thymic) large B cell lymphoma treated with second-line therapy and intent to transplant. Biol Blood Marrow Transplant 24:2133-2138, 2018.

3. Zinzani PL, Pellegrini C, Chiappella A, et al: Brentuximab vedotin in relapsed primary mediastinal large B-cell lymphoma: Results from a phase 2 clinical trial. Blood 129:2328-2330, 2017.

4. Armand P, Rodig S, Melnichenko V, et al: Pembrolizumab in relapsed or refractory primary mediastinal large B-cell lymphoma. J Clin Oncol 37:3291-3299, 2019.

5. Ansell SM, Minnema MC, Johnson P, et al: Nivolumab for relapsed/refractory diffuse large B-cell lymphoma in patients ineligible for or having failed autologous transplantation: A single-arm, phase II study. J Clin Oncol 37:481-489, 2019.

6. Godfrey J, Tumuluru S, Bao R, et al: PD-L1 gene alterations identify a subset of diffuse large B-cell lymphoma harboring a T-cell-inflamed phenotype. Blood 133:2279-2290, 2019.


Related Articles

Pembrolizumab in Relapsed or Refractory Primary Mediastinal Large B-Cell Lymphoma: KEYNOTE-013 and KEYNOTE-170

Findings in the phase Ib KEYNOTE-013 and phase II KEYNOTE-170 trials, reported in the Journal of Clinical Oncology by Philippe Armand, MD, PhD, of Dana­Farber Cancer Institute, and colleagues, indicate that pembrolizumab is active in relapsed or refractory primary mediastinal large B-cell...