A prognostic model based on six clinical factors can help predict overall survival (OS) in patients with advanced urothelial cancer (UC) who are treated with the anti-programmed cell death ligand 1 (PD-L1) immune checkpoint inhibitor atezolizumab (Tecentriq, Roche) after platinum-based therapy.

The US Food and Drug Administration has approved five new immunotherapies for advanced UC that has progressed during or after platinum-based chemotherapy, including atezolizumab.

"Based on the rapid availability of new therapies, we thought it was important to try to assess which patients might benefit the most from atezolizumab, which is one type of these new therapies," lead investigator, Gregory Pond. "We believe we've developed the first prognostic model that, once confirmed in larger studies, could provide a critical decision-making tool for clinicians."

To develop the model, the researchers carried out a post hoc analysis of patient-level data from two prospective trials involving patients with advanced UC treated with atezolizumab in the postplatinum setting. The phase 2 IMvigor210 study of 310 patients served as the training data set, and the phase 1a PCD4989g study of 95 patients served as the validation data set.

The researchers considered various clinical factors that had been previously shown to predict survival in patients with advanced UC receiving chemotherapy.

The "optimal" prognostic model for OS consisted of six factors:

  1. ECOG-PS of 1 or greater vs 0 (hazard ratio [HR], 1.64; 95% confidence interval [CI], 1.20 – 2.24; = .002);

  2. Presence of liver metastasis (HR, 1.45; 95% CI, 1.08 – 1.94; = .014);

  3. Elevated platelet count (≥400 × 10LHR, 1.73; 95% CI, 1.14 – 2.61; = .010);

  4. Increased neutrophil-to-lymphocyte ratio of 5 or greater (HR, 1.84; 95% CI, 1.45 – 2.34; < .001);

  5. Elevated lactate dehydrogenase level of 280 U/L or greater (HR, 1.54; 95% CI, 1.19 – 1.99; =  .001); and

  6. Anemia (hemoglobin level < 10 g/dL; HR, 1.60; 95% CI, 1.17 – 2.21; = .004).

In the IMvigor210 dataset, median OS was 19.6 months for patients with zero or one factor, 5.9 months for patients with two to three factors, and 2.8 months for those with four or more. In the PCD4989g data set, corresponding median OSs were 19.4, 7.2, and 2.6 months.

"The initial results of our study are very promising in both the training and validation data sets," Dr Pond said. "The model does require further evaluation and further refinements. For example, we need to look at how the model performs in data sets of larger sample size, and we also want to see how it works with other checkpoint inhibitors," he added.

Sumanta K. Pal, MD, noted that the introduction of immunotherapy represents a "breakthrough" in bladder cancer management, but it's important to bear in mind that only about one quarter of patients will have substantial tumor shrinkage with these therapies, and only a relatively small number of patients will have extended survival with these therapies.

"Until the results of this study, there was no way to easily discern prognosis and identify who might stand to benefit most," Dr Pal said. "This easily applied score developed by Dr Pond et al. based on parameters readily available in the patient's chart provides tremendous input."

"While I would not necessarily withhold therapy on the basis of an anticipated poor prognosis, I would consider using this information in counseling patients who want to be better informed of potential outcomes with immunotherapy," Dr Pal said.

Dr Pal also noted that multiple ongoing studies are testing combinations of chemotherapy with immunotherapy in advanced UC. "If these studies are positive, there would be a massive paradigm shift in how we manage advanced bladder cancer. At that point in time, we would have to determine if the model established by Dr Pond would remain relevant in that climate," Dr Pal said.