Prognostic model of all-cause mortality at 30 days in patients with cancer and COVID-19

Bookmark and Share
Published: 11 Sep 2022
Views: 16
Rating:
Save
Dr Chris Labaki - Dana-Farber Cancer Institute, Boston, USA

Dr Chris Labaki speaks to ecancer about a new prognostic model of all-cause mortality at 30 days in patients with cancer and COVID-19.

Cancer patients are at increased risk of dying from COVID-19. Known risk factors for 30-day all-cause mortality (ACM-30) in patients with cancer can be older age, sex, smoking status, performance status, obesity, and co-morbidities. Prof Halabi, the primary author of this poster, and her team, hypothesised that common clinical and laboratory parameters would predict a higher risk of ACM at 30 days and that a machine learning approach (random forest) could produce high accuracy.

The results of this study found the median age of COVID-19 diagnosis to be 65 years. Over half were never smokers and the median body mass index was 28.2. This prognostic model based on readily available clinical and laboratory values can be used to estimate individual survival probability within 30-days for COVID-19.

This study was conducted with the COVID-19 and Cancer Consortium. There have been a lot of data, small retrospective studies, discussing or evaluating the influence of immunotherapy on the outcomes of patients with cancer that develop COVID-19. Looking at the biology of COVID-19, it has been associated sometimes with a multi-inflammatory syndrome, so an over-activation of the immune system itself. Also in parallel, COVID-19 has been shown to be associated sometimes with a state of immunosuppression that would be technically associated with worse outcomes. 

So regarding the complex biology of COVID-19 and looking specifically at patients with cancer who represent, one, a very vulnerable population in relation to COVID-19 as compared to the general healthy population, and, two, who receive frequently immunotherapy as part of their treatment regimen, we thought to evaluate the clinical outcomes of patients with cancer and COVID-19 in relation to immunotherapy as an exposure but also immunosuppression. We tried to look at what is the true influence of immunotherapy and immunosuppression on the outcomes of patients with cancer and COVID-19.

We used data from the COVID-19 and Cancer Consortium, which I am part of and our team at Dana-Farber was part of. We led a project where we evaluated how does immunosuppression or immunotherapy interact with clinical outcomes. The primary endpoint was COVID-19 severity, however, in relation to the inflammatory syndrome that is usually seen in patients with COVID-19, not very frequently but reported at least in the literature, we also defined this secondary endpoint being the cytokine storm. We defined it based on a set of both biological and clinical parameters to encompass the overall picture of a true cytokine storm which should be technically related to an over-activation of the immune system. So these were our two endpoints. 

Now, we controlled for a large set of clinical and demographic parameters, including vaccination status, all demographic parameters, previous therapies received. Our two exposures of interest were recent anti-cancer systemic therapies, that’s one, further stratified into IO, or immunotherapy, based or non-IO systemic therapies and immunosuppression status. So we defined immunosuppression at baseline, meaning before COVID-19. To account for a potential interaction between the two exposures, being immunosuppression and immunotherapy, we included an interaction term between them in the multivariable analysis that we conducted.

What we found is for both endpoints, COVID-19 severity and incidence of cytokine storm, we found a significant interaction, meaning that we would have to stratify by immunosuppression status to appreciate the true influence of immunosuppression and immunotherapy in relation to the outcomes of patients with cancer and COVID-19. What we saw overall, the final results are that in the subgroup of patients who have baseline immunosuppression the recent administration, before COVID-19, of IO anti-cancer therapy, so immunotherapy, and non-IO systemic anti-cancer therapy was associated with worse outcomes as compared to patients with cancer and immunosuppression who developed COVID-19 but did not receive any recent systemic therapy. 

However, in patients with cancer and COVID-19 who did not have immunosuppression, the recent administration of any kind of systemic anti-cancer therapy was not associated with worse outcomes as compared to untreated patients. Meaning that you have to combine two conditions in order to relate to worse outcomes among patients with cancer and COVID-19, these two conditions being baseline immunosuppression and recently received systemic anti-cancer therapies.

So this would help achieve a clarification for the entire field because there have been so far many conflicting data. So some studies would infer worse outcomes in relation to immunotherapy; other studies would refer to non-difference in outcomes, so a non-influence of immunotherapy in relation to clinical outcomes. So our study helps to better define the specific subgroups of patients with cancer using a very large dataset of 12,000 patients with cancer and COVID-19 and encompassing time periods from March 2020 to May 2022, meaning that we cover all the different phases of the pandemic and accounting for a very comprehensive set of both demographic and clinical variables. This helps to better define the specific subgroups of patients with cancer and COVID-19 who might be very vulnerable to worse outcomes.