Behind-the-paper: Whole-genome Sequencing Identifies Responders to Pembrolizumab in Relapse/Refractory Natural-killer/T Cell Lymphoma

The outlook for patients with relapse/refractory Natural-killer/T cell lymphoma (NKTCL) is bleak. However, recent studies have shown dramatic response rates to pembrolizumab for such patients. Thus, this motivated us to search for molecular biomarkers of response to pembrolizumab in NKTCL.

Like Comment
Read the paper

Our proband patient (NKTL1), a 52-year-old Chinese male, was first diagnosed with NKTCL in November 2013. He achieved complete response with 4 cycles of the GELOX (gemcitabine, oxaliplatin, and L-asparaginase) chemotherapy regime, but suffered a relapse within 2 years (Fig. 1).  

Fig 1. Treatment timeline for proband patient NKTL1.

In April 2016, after failing an additional 4 lines of salvage therapies, his condition took a nose-dive with evidence of multi-organ failure and severe pancytopenia. As his high-risk clinical condition precluded him from enrolling in any clinical trial, an off-label infusion of pembrolizumab at 200 mg was initiated in May 2016. Interestingly, 1 week after the infusion, he began to show gradual clinical improvement and 2 weeks subsequently, his liver derangements were completely normalized, and his pancytopenia had improved remarkably. The patient achieved complete response with PET-CT scans showing undetectable hypermetabolic foci 8 weeks after the first administered dose of pembrolizumab (Fig. 2). It is heartening to learn that the patient is still in complete remission at the time of this writing.  

Fig 2. PET-CT scans of patient NKTL1 before and after pembrolizumab.

As the observed complete response to pembrolizumab in our proband patient was unheard of in relapsed/refractory NKTCL, we were immensely interested in the molecular profile of this tumor. We whole-genome sequenced the pre-pembrolizumab treated tumor and buccal swab DNA from this patient and found a somatic micro-inversion in the 3’ UTR of the PD-L1 gene. Knowing that the canonical suppressive function of the 3’UTR could have been disrupted and upregulated PD-L1, we contacted our collaborators for more pre-pembrolizumab treated NKTCL samples. To our surprise, structural rearrangements to the 3’UTR of PD-L1 is common (23.8%, 5/21) in NKTCL and are enriched within only the responding patients with NKTCL to pembrolizumab. Lastly, I would like to thank our patients for making this study possible and hope that this study will potentiate better outcomes for future patients. 

Details of our findings can now be found in the published version of the manuscript -> here.

Jing Quan Lim

Research Fellow, National Cancer Centre Singapore