Predicting Outcomes of COVID-19 in Patients With Cancer

Findings from several studies can help clinicians predict outcomes of COVID-19 in patients with cancer, according to a recent European Society for Medical Oncology (ESMO) webinar presentation.¹

“Cancer has been consistently confirmed as a risk factor for more severe outcomes in patients with COVID-19,” said presenter Luís Castelo-Branco, MD, PharmD, of the Algarve Medical and University Centre in Faro, Portugal. 

Dr Castelo-Branco cited a systematic review of 207 studies suggesting that, compared with the general population, cancer patients have more than twice the risk of severe COVID-19 and a 35% higher risk of death.²

Dr Castelo-Branco went on to highlight studies suggesting that certain risk factors are associated with outcomes of COVID-19 in patients with cancer, and artificial intelligence (AI) tools can use those risk factors to predict outcomes.

Risk Factors Tied to Outcomes

A COVID-19 and Cancer Consortium (CCC19) study of 4966 patients showed that patients with active cancer have worse outcomes of COVID-19, with higher mortality rates observed in active (22%) or progressing cancers (34%).³

Other risk factors for worse outcomes in the CCC19 study included older age, male sex, obesity, Black race, Hispanic ethnicity, cardiovascular disease, pulmonary disease, renal disease, and diabetes. 

Cancer-related risk factors included recent cytotoxic chemotherapy, hematologic malignancy, and treatment with R-CHOP (rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisone), platinum-based therapy combined with etoposide, and DNA methyltransferase inhibitors. 

Some of these risk factors were observed in data from the ESMO CoCARE registry as well.⁴ In this study, risk factors associated with severe COVID-19 were male sex, older age, non-Caucasian ethnicity, ECOG performance status of 2 or higher, body mass index of 25 or higher, 1 or more comorbidities, hematologic cancer, progressive disease, and a lack of COVID-19 symptoms.

In the OnCOVID study, researchers found that systemic inflammation is associated with COVID-19 outcomes.⁵ The neutrophil-to-lymphocyte ratio, OnCOVID Inflammatory Score (formerly Prognostic Nutritional Index), and modified Glasgow Prognostic Score all worsened at COVID-19 diagnosis and were associated with inferior survival. 

AI May Predict Outcomes

Studies have also suggested that AI tools can predict disease severity and death in patients with cancer and COVID-19.^6-8

In one study,⁶ random forest modeling suggested the most significant predictors of 30-day death after COVID-19 (in decreasing order) were:

• Older age
• Treatment of advanced or metastatic disease
• Tumor type (with the highest mortality rates seen in respiratory tract, brain, and unknown primary cancers)
• COVID-19-related symptom burden at baseline evaluation
• Treatment regimen (with higher mortality rates seen in patients receiving immunotherapy combinations). 

This model predicted outcomes with 89% sensitivity and 88% specificity.

In another study,⁷ researchers used clinical variables collected on or before a patient’s COVID-19 diagnosis date to develop an algorithm that would classify the patients into 3 outcome categories: 

• Severe-early, defined as requiring high levels of oxygen support within 3 days of testing positive for SARS-CoV-2
• Severe-late, defined as requiring high levels of oxygen after 3 days
• Non-severe, defined as never requiring oxygen support.

The algorithm classified patients into these categories with an area under the receiver operating characteristic curve ranging from 70% to 85%.

Yet another AI tool, “CORONET,” has been shown to predict which patients with cancer and COVID-19 will require hospital admission.⁸ In a cohort of 672 patients, CORONET recommended admission for 96% of patients who required oxygen and 99% of patients who ultimately died.

CORONET is freely available online.⁹ According to the website, the tool asks for details about the patient, their cancer, and blood test results on presentation to the hospital with symptoms of COVID-19. The tool uses data about the admission, requirement for oxygen, and survival of similar patients to predict the likely outcome of the patient in question. 

Despite the aforementioned discoveries and advances, Dr Castelo-Branco said bigger and more powerful data analyses are needed. More data are required to better understand the influence of COVID-19 vaccination and different SARS-CoV-2 variants on outcomes in patients with cancer.  

References

1. Castelo-Branco L. Risk prediction for COVID related morbidity and mortality in patients with cancer: Available clinical and laboratory algorithms. ESMO webinar. Aired February 16, 2022. Accessed March 10, 2022.

2. Izcovich A, Ragusa MA, Tortosa F, et al. Prognostic factors for severity and mortality in patients infected with COVID-19: A systematic review. PloS One. 2020;15(11):e0241955. doi:10.1371/journal.pone.0241955 

3. Grivas P, Khaki AR, Wise-Draper TM, et al. Association of clinical factors and recent anticancer therapy with COVID-19 severity among patients with cancer: A report from the COVID-19 and Cancer Consortium. Ann Oncol. 2021;32(6):787-800. doi:10.1016/j.annonc.2021.02.024 

4. Romano E, Gennatas S, Rogado J, et al. 1567MO COVID-19 and cancer: First report of the ESMO international, registry-based, cohort study (ESMO CoCARE). Ann Oncol. 2021;32 (suppl 5): S1133. doi:10.1016/j.annonc.2021.08.1560

5. Dettorre GM, Dolly S, Loizidou A, et al. Systemic pro-inflammatory response identifies patients with cancer with adverse outcomes from SARS-CoV-2 infection: The OnCovid Inflammatory Score. J Immunother Cancer. 2021 Mar;9(3):e002277. doi:10.1136/jitc-2020-002277

6. Dienstmann R, Menezes M, Costa e Silva M, et al. Machine learning prediction of COVID-19 mortality in cancer patients. J Clin Oncol. 2021; 39(15_suppl):1558-1558. doi:10.1200/JCO.2021.39.15_suppl.1558

7. Navlakha S, Morjaria S, Perez-Johnston R, et al. Projecting COVID-19 disease severity in cancer patients using purposefully-designed machine learning. BMC Infect Dis. 2021; 21;391. https://doi.org/10.1186/s12879-021-06038-2

8. Lee R, Wysocki O, Zhou C, et al. CORONET; COVID-19 in Oncology evaluatiON Tool: Use of machine learning to inform management of COVID-19 in patients with cancer. J Clin Oncol. 2021;39(15_suppl), 1505-1505. doi:10.1200/JCO.2021.39.15_suppl.1505

9. CORONET: COVID-19 Risk in ONcology Evaluation Tool. Accessed March 10, 2022. https://coronet.manchester.ac.uk/