Model May Predict Post-treatment Hodgkin Lymphoma Outcomes

Liam Davenport

December 11, 2020

Combining positron emission tomography (PET) with tumour gene expression profiling straight after treatment may be able to identify which patients with early stage classical Hodgkin lymphoma (cHL) are likely to experience disease progression or another event during follow-up, suggests an analysis of UK trial data.

The research was presented at the virtual 62nd American Society of Hematology (ASH) Annual Meeting and Exposition on December 7.

'Promising Utility'

The study was a sub-analysis of the UK National Cancer Research Institute's RAPID clinical trial, which examined whether patients with a negative PET scan after three cycles of doxorubicin hydrochloride, bleomycin sulfate, vinblastine, and dacarbazine (ABVD) chemotherapy needed consolidation radiotherapy to delay or prevent disease progression.

For the current analysis, tumour samples from more than 220 patients from RAPID underwent gene expression profiling.

This identified three genes significantly associated with the risk of a cHL event during follow-up: the two immune cell-related genes IL15RA and CD22, and the apoptosis-related gene BID.

A model combining the three genes with interim PET scores was significantly predictive of patients at high versus low risk of a cHL-related event during follow-up, with a high score associated with a more than eightfold increased risk of an event.

Presenting the findings, Dr Kim Linton, University of Manchester and The Christie NHS Foundation Trust, Manchester, said the model showed "promising utility for enhanced prediction of cHL event-free survival" in early stage disease, independently of widely used pre-treatment clinical risk scores.

She believes the findings "warrant further evaluation in an independent cohort", with the aim of "improving the precision of individualised treatment to improve disease control and minimise late toxicity".

Dr Linton added the results also validate previous reports indicating that overexpression of the B-cell marker CD22 has a protective function in cHL, thus "conferring a more favourable outcome".

'Interesting, Important'

Dr Catherine Diefenbach, director of the clinical lymphoma program at New York University Langone's Perlmutter Cancer Center, who was not involved in the study, spoke to Medscape News UK.

She said the study is "interesting, and I think risk stratification in Hodgkin's lymphoma is really important, especially because...we really want to spare the patients who are going to do well as much treatment as possible so we spare them long-term side effects".

She continued that clinicians also "want to not minimise but treat to cure the patients who are at the highest risk of relapse, so that's really our goal with Hodgkin lymphoma patients".

Dr Diefenbach said that the current study nevertheless has a couple of caveats, the "biggest one" being that "this is a non-validated study", in that the analysis was run on the full dataset with no independent validation in a separate cohort of patients.

Moreover, "there are so few events" during the follow-up period, "which is a good thing for patients but it's hard to know the significance" of the model "with this small number of events".

The results therefore require confirmation in further studies before the model can be said to be "ready for prime time".

In terms of its potential role in relation to other risk stratification scores, Dr Diefenbach, who designed the IPS-3 score for advanced Hodgkin lymphoma, said the "whole goal with these clinical risk scores is to be able to combine them with a biologic risk score so that you can get enhanced risk profiling, and really personalised medicine".

She added that they are "not competing with each other", and the current model is distinguished from most other risk scores due to being designed for post-treatment assessment, which could "tell you for example, if it was validated, if you need to augment therapy or get away with doing less for a particular patient".

Study Details

Dr Linton began her presentation by noting that, while combined modality treatment is the standard of care for early stage cHL, long-term survival is undermined by late toxicities, resulting in patients, over time, being twice as likely ultimately to die from treatment toxicity as from the disease itself.

Although a negative interim PET scan after up to three cycles of ABVD is able to predict a low relapse rate in patients with non-bulky stage 1a/2a cHL, several studies have shown that it is not, on its own, able to select patients suitable for omission of radiotherapy.

To determine whether gene expression profiles in diagnostic tumour tissue could improve patient selection, the researchers conducted BioPET, a sub-analysis of the RAPID trial.

For RAPID, 571 cHL patients with stage 1a/2a disease and no mediastinal bulk were treated with three cycles of ABVD.

The 426 patients who subsequently had a complete metabolic response score of 1 or 2 were then randomised to 30 Gy of involved-field radiation therapy (IFRT) or no further therapy. The 145 patients who had a positive PET, defined as a score of 3–5, were given a fourth cycle of ABVD and then IFRT.

Patients were included in the BioPET analysis if they had undergone a pre-treatment European Organisation for Research and Treatment of Cancer (EORTC) and German Hodgkin Study Group (GHSG) risk assessment and had 3 years of event-free survival data following treatment.

To conduct the gene profiling, they selected 57 candidate gene biomarkers from dysregulated pathways in cHL, and used a standard assay (Quantigene Plex 2.0, ThermoFisher Scientific) to assess each patient's gene expression profile.

They were able to examine 227 patient samples from the RAPID trial, alongside 15 RNA reference controls and 12 formalin-fixed, paraffin-embedded control tissue samples.

The median age of the patients was 36 years, 48.0% were female and 71.4% had stage 2A disease. The pretreatment risk scores were favourable in approximately 66% of patients, and all three treatment arms were well represented.

After a median follow-up of 55 months, there were 21 cHL events.

After adjusting for treatment group, PRF1 and BCL2L1 genes were significantly associated with an interim PET score of 3–5 on multivariate analysis, at odds ratios of 1.49 and 0.65, respectively (p=0.03 for both).

Three genes were also significantly associated with a cHL event on multivariate analysis: IL15RA, at a hazard ratio of 0.33 (p=0.02); CD22, at a hazard ratio of 0.54 (p<0.001); and BID, at a hazard ratio of 3.66 (p<0.001).

Combining CD22, BID and IL15RA expression with interim PET scores in a risk model, the team found that, at a cut-off score of 0.69, it was able to identify high versus low risk patients at a hazard ratio of 8.85 (p<0.001).

They determined that the true positive rate with the model was 67.7%, while the false positive rate was 9.5%.

Dr Linton added that, in a separate multivariate model, CD22 was a significant predictor of treatment failure, defined as a PET score of 4–5 or a cHL event.

No funding declared.

Linton declares Roche: Consultancy; Celgene: Membership on an entity's Board of Directors or advisory committees; Janssen: Other: Travel, accommodations, expenses ; Celgene: Other: Travel, accommodations, expenses; Beigene: Membership on an entity's Board of Directors or advisory committees; Takeda: Honoraria; Gilead: Membership on an entity's Board of Directors or advisory committees; Karyopharm: Membership on an entity's Board of Directors or advisory committees.

62nd ASH Annual Meeting and Exposition: Abstract 2918. Presented 7 December.

Blood 2020; 136 (Supplement 1): 18–19. doi: 10.1182/blood-2020-136318


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: