Characteristics and clinical subtypes of cancer patients in the intensive care unit: a retrospective observational study for two large databases

Outcome prediction for critical care patients with respiratory neoplasms using a multilayer perceptron neural network

OBJECTIVE

The variation in mortality rates of intensive care unit oncological patients may imply that clinical characteristics and prognoses are very different between specific subsets of patients with cancer. The specific characteristics of patients with cancer have not been included as risk factors in the established severity-of-illness scoring systems and comorbidity scores, showing limitations in predicting mortality risk. This study aimed to devise a predictive tool for in-hospital mortality for adult patients with a respiratory neoplasm admitted to the intensive care unit, using an artificial neural network.

METHODS

A total of 1,221 stays in the intensive care unit from the Beth Israel Deaconess Medical Center were studied. The primary endpoint was the all-cause in-hospital mortality prediction. An artificial neural network was developed and compared with six severity-of-illness scores and one comorbidity score. Model building was based on important predictors of lung cancer mortality, such as several laboratory parameters, demographic parameters, organ-supporting treatments, and other clinical information. Discrimination and calibration were assessed.

RESULTS

The AUROC for the multilayer perceptron was 0.885, while it was <0.74 for the conventional systems. The AUPRC for the multilayer perceptron was 0.731, whereas it was ≤0.482 for the conventional systems. The superiority of multilayer perceptron was statistically significant for all pairwise AUROC and AUPRC comparisons. The Brier Score was better for the multilayer perceptron (0.109) than for OASIS (0.148), SAPS III (0.163), and SAPS II (0.154).

CONCLUSION

Discrimination was excellent for multilayer perceptron, which may be a valuable tool for assessing critically ill patients with lung cancer.

A Systematic Review and Meta-Analysis Evaluating Geographical Variation in Outcomes of Cancer Patients Treated in ICUs

The reported mortality rates of cancer patients admitted to ICUs vary widely. In addition, there are no studies that examined the outcomes of critically ill cancer patients based on the geographical regions. Therefore, we aimed to evaluate the mortality rates among critically ill cancer patients and provide a comparison based on geography.

DATA SOURCES

PubMed, EMBASE, and Web of Science.

STUDY SELECTION

We included observational studies evaluating adult patients with cancer treated in ICUs. We excluded non-English studies, those with greater than 30% hematopoietic stem cell transplant or postsurgical patients, and those that evaluated a specific type of critical illness, stage of malignancy, or age group.

DATA EXTRACTION

Two reviewers independently applied eligibility criteria, assessed quality, and extracted data. Studies were classified based on the continent in which they were conducted. Primary outcomes were ICU and hospital mortality. We pooled effect sizes by geographical region.

DATA SYNTHESIS

Forty-six studies were included (n = 110,366). The overall quality of studies was moderate. Most of the published literature was from Europe (n = 22), followed by North America (n = 9), Asia (n = 8), South America (n = 5), and Oceania (n = 2). Pooled ICU mortality rate was 38% (95% CI, 33-43%); the lowest mortality rate was in Oceania (26%; 95% CI, 22-30%) and highest in Asia (51%; 95% CI, 44-57%). Pooled hospital mortality rate was 45% (95% CI, 41-49%), with the lowest in North America (37%; 95% CI, 31-43%) and highest in Asia (54%; 95% CI, 37-71%).

CONCLUSIONS

More than half of cancer patients admitted to ICUs survived hospitalization. However, there was wide variability in the mortality rates, as well as the number of available studies among geographical regions. This variability suggests an opportunity to improve outcomes worldwide, through optimizing practice and research.