Findings presented at the European Society for Medical Oncology’s (ESMO) Asia Virtual Congress 2020, held November 20 – 22, 2020, demonstrate that several key biomarkers of inflammation may identify cancer patients at increased risk of mortality from COVID-19. [1]

In the study, Gino Dettorre, MRes, of the Department of Surgery and Cancer, Imperial College London – Hammersmith Hospital in London, UK, and colleagues from 19 European centers, tested and validated their hypothesis. [1].

With much uncertainty as to the contribution of cancer patients’ features on severity and mortality from Covid-19 and little guidance as to the role of anti-cancer and anti-Covid-19 therapy, the investigators wanted to understand the risks in this population. They found a significant association between key biomarkers of inflammation and decreased Overall Survival (OS).[1][2]

The investigators, prompted by the understanding that systemic inflammation is common to both cancer progression and SARS-CoV-2 infection, wanted to find out whether inflammatory biomarkers can identify cancer patients with COVID-19 who are at risk of poorer outcomes.[1][2]

To confirm their hypothesis, the scientists retrospectively accrued 1,318 consecutive cancer patients infected with SARS-CoV-2 from February 27, 2020 to June 23, 2020 at 23 academic centers in the United Kingdom, Spain, Italy, Germany, and Belgium. Eligible patients were aged ≥18 and those with leukemia, myeloma, or insufficient data were excluded from the study.

Advertisement #3

OnCovid study
In the OnCovid-study, a multi-center retrospective observational study sponsored by Imperial College London, investigators evaluated neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), prognostic nutritional index (PNI), modified Glasgow prognostic score (mGPS), and prognostic index (PI) as prognostic biomarkers for the increased risk. The NLR, PLR, and PNI were dichotomized around their medians. The validation set confirmed prognostic biomarkers evaluated in the training set

Of 1,071 patients, 529 were sorted into a training set and 542 into a validation set. In the training and validation sets, respectively, patients were matched by age (67.9 ±13.3 versus 68.5 ±13.5), active malignancy at COVID-19 diagnosis (66.7% versus 61.6%), presence of >1 comorbidity (52.1% versus 49.8%), and prevalence of complications, including respiratory failure (58.0% versus 59.0%) and acute respiratory distress syndrome (ARDS; 11.5% versus 12.9%).

Assessment of biomarkers in the training set showed that higher mortality rates were associated with NLR>6 (44.6% versus 28%; p < 0.0001), PNI<40 (46.6% versus 20.9%; p < 0.0001), mGPS (50.6% for mGPS2 versus 30.4% and 11.4% for mGPS1 and 0; p < 0.0001), and PI (50% for PI2 versus 40% for PI1 and 9.1% for PI0; p < 0.0001).

Following multivariable hazard ratio calculations and Harrell’s C-index computation, patients with a poor risk PNI, a computation derived from hypoalbuminemia and lymphocytopenia, proved most susceptible to severe COVID-19. The mGPS (calculated via C-reactive protein levels and hypoalbuminemia), PI (calculated via C-reactive protein levels and leukocytosis), and NLR (the ratio of neutrophils to lymphocytes) emerged as most to least predictive of fatal COVID-19 respectively. © Gino Dettorr

The findings were confirmed in the validation set (p < 0.001). The investigators found that patients in poor-risk categories had a significantly shorter median OS. Specifically, those with NLR>6 had a median OS of 30 days (95% confidence interval [CI] 1-63), those with PNI<40 had a median OS of 23 days (95% CI 10-35), mGPS2 corresponded with median OS of 20 days (95% CI 8-32), and patients with PI2 demonstrated a median OS of 23 days (95% CI 1-56) compared to patients in good risk categories wherein median OS was not reached (p < 0.001 for all comparisons). No association between PLR and survival was found.

Analyses of survival in the validation set confirmed NLR (p < 0.0001), PNI (p < 0.0001), PI (p < 0.01) and mGPS (p < 0.001) as predictors of survival.

In the study, inflammatory markers’ hazard ratios (HRs) were compared in a multivariable Cox regression model with conditional backward elimination against each other and against previously established prognostic factors for COVID-19 outcome including sex, age, comorbid burden, malignancy status, and receipt of anti-cancer therapy at COVID-19 diagnosis.

Among both inflammatory markers of interest and previously established prognostic factors, the PNI was the only factor to emerge with a statistically significant HR in both training (HR 1.97; 95% CI 1.19-3.26, p = 0.008) and validation set (HR 2.48; 95% CI 1.47-4.20, p = 0.001) analysis. The order of inflammatory indices’ prognostic relevance was determined via Harrell’s C-index.

Study limitations
The investigators acknowledge that the study was limited to secondary use of information previously collected in the course of normal care (without an intention to use it for research at the time of collection).

However, keeping in mind these limitations, the investigators believe that their findings clearly underscore that systemic inflammation is a key driver of mortality from SARS-CoV-2 in cancer patients, and the NLR, PNI, mGPS, and PI are externally validated biomarkers that quantify systemic inflammation in patients with cancer and can be used as bedside tests to stratify patients at risk of poorer outcome from COVID-19, with an emphasis on hypoalbuminemia and lymphocytopenia as computed by the PNI.

Clinical trials
COVID-19 and Cancer Patients (OnCovid) – NCT04393974

[1] Dettorre G, Diamantis N, Loizidou A, et al. The systemic pro-inflammatory response identifies cancer patients with adverse outcomes from SARS-CoV-2 infection. ESMO Asia Virtual Congress 2020 (November 20 – 22, 2020). Abstract 319O. Annals of Oncology (2020) 31 (suppl_6): S1366-S1370. 10.1016/annonc/annonc363
[2] Pinato DJ, Zambelli A, Bertuzzi A, Marrari A, Saoudi-Gonzalez N, Mirallas O, Galazi M, et al. Clinical portrait of the SARS-CoV-2 epidemic in European cancer patients. ESMO Asia Virtual Congress 2020 (November 20 – 22, 2020).Abstract 318O Annals of Oncology (2020) 31 (suppl_6): S1366-S1370. 10.1016/annonc/annonc363

Featured image: SARVS-CoV-2. Photo courtesy: CDC | Centers for Disease Control and Prevention. Used with permission.

Advertisement #5