Implementation of COVID-19 measures may have induced concerns about access and quality of health care for cancer patients with bone metastases, and it may have affected their quality of life. In this study, we evaluated the effect of the first COVID-19 lockdown on quality of life and emotional functioning of patients with stage IV cancer treated for painful bone metastases in the UMC Utrecht, the Netherlands. A COVID-19 specific questionnaire was sent to active participants in the Prospective Evaluation of interventional StudiEs on boNe meTastases (PRESENT) cohort, consisting of patients irradiated for metastatic bone disease. Patient reported outcomes (PROs) were compared with the last two PROs collected within the PRESENT cohort before the COVID-19 lockdown in the Netherlands on the 16th of March. For the 169 (53%) responders, median age at start of lockdown was 68 years (range 38–92) and 62% were male. Patients reported a statistically significant decrease in emotional functioning (83.6 to 79.2, P = 0.004) and in general quality of life score during the COVID-19 lockdown (72.4 to 68.7, P = 0.007). A steep increase in feeling isolated was reported (18% before and 67% during lockdown). This study has shown a strong increase in the experience of isolation and a decrease of emotional functioning and general quality of life during the COVID-19 lockdown in cancer patients with bone metastases. Due to the nature of the treatment of this patient population, efforts should be made to minimize these changes during future lockdowns.

Deep learning has shown remarkable results for image analysis and is expected to aid individual treatment decisions in health care. Treatment recommendations are predictions with an inherently causal interpretation. To use deep learning for these applications in the setting of observational data, deep learning methods must be made compatible with the required causal assumptions. We present a scenario with real-world medical images (CT-scans of lung cancer) and simulated outcome data. Through the data simulation scheme, the images contain two distinct factors of variation that are associated with survival, but represent a collider (tumor size) and a prognostic factor (tumor heterogeneity), respectively. When a deep network would use all the information available in the image to predict survival, it would condition on the collider and thereby introduce bias in the estimation of the treatment effect. We show that when this collider can be quantified, unbiased individual prognosis predictions are attainable with deep learning. This is achieved by (1) setting a dual task for the network to predict both the outcome and the collider and (2) enforcing a form of linear independence of the activation distributions of the last layer. Our method provides an example of combining deep learning and structural causal models to achieve unbiased individual prognosis predictions. Extensions of machine learning methods for applications to causal questions are required to attain the long-standing goal of personalized medicine supported by artificial intelligence.

No abstract

No abstract