Prognostic Factors of Malignant Pleural Effusion among Palliative Care Outpatients: A Retrospective Study

Automated Detection, Segmentation, and Classification of Pleural Effusion From Computed Tomography Scans Using Machine Learning

OBJECTIVE

This study trained and evaluated algorithms to detect, segment, and classify simple and complex pleural effusions on computed tomography (CT) scans.

MATERIALS AND METHODS

For detection and segmentation, we randomly selected 160 chest CT scans out of all consecutive patients (January 2016-January 2021, n = 2659) with reported pleural effusion. Effusions were manually segmented and a negative cohort of chest CTs from 160 patients without effusions was added. A deep convolutional neural network (nnU-Net) was trained and cross-validated (n = 224; 70%) for segmentation and tested on a separate subset (n = 96; 30%) with the same distribution of reported pleural complexity features as in the training cohort (eg, hyperdense fluid, gas, pleural thickening and loculation). On a separate consecutive cohort with a high prevalence of pleural complexity features (n = 335), a random forest model was implemented for classification of segmented effusions with Hounsfield unit thresholds, density distribution, and radiomics-based features as input. As performance measures, sensitivity, specificity, and area under the curves (AUCs) for detection/classifier evaluation (per-case level) and Dice coefficient and volume analysis for the segmentation task were used.

RESULTS

Sensitivity and specificity for detection of effusion were excellent at 0.99 and 0.98, respectively (n = 96; AUC, 0.996, test data). Segmentation was robust (median Dice, 0.89; median absolute volume difference, 13 mL), irrespective of size, complexity, or contrast phase. The sensitivity, specificity, and AUC for classification in simple versus complex effusions were 0.67, 0.75, and 0.77, respectively.

CONCLUSION

Using a dataset with different degrees of complexity, a robust model was developed for the detection, segmentation, and classification of effusion subtypes. The algorithms are openly available at https://github.com/usb-radiology/pleuraleffusion.git.

Predicting the survival in patients with malignant pleural effusion undergoing indwelling pleural catheter insertion

CONTEXT

Malignant pleural effusion (MPE) is a common comorbid condition in advanced malignancies with variable survival.

AIMS

The aim of this study was to predict the survival in patients with MPE undergoing indwelling pleural catheter (IPC) insertion.

SETTINGS AND DESIGN

This was a cross-sectional study conducted at Shaukat Khanum Memorial Cancer Hospital and Research Centre, Lahore, Pakistan.

METHODS

One hundred and ten patients with MPE who underwent IPC insertion from January 2011 to December 2019 were reviewed. Kaplan-Meier method was used to determine the overall survival (OS) of the patient's cohort with respect to LENT score.

STATISTICAL ANALYSIS USED

The IBM SPSS version 20 was used for statistical analysis.

RESULTS

We retrospectively reviewed 110 patients who underwent IPC insertion for MPE, with a mean age of 49 ± 15 years. 76 (69.1%) patients were females, of which majority 59 (53.6%) had a primary diagnosis of breast cancer. The LENT score was used for risk stratification, and Kaplan-Meier survival curves were used to predict the OS. The proportion of patients with low-risk LENT score had 91%, 58%, and 29% survival, the moderate-risk group had 76%, 52%, and 14% survival, and in the high-risk group, 61%, 15%, and 0% patients survived at 1, 3, and 6 months, respectively. In addition, there was a statistically significant survival difference (P = 0.05) in patients who received chemotherapy pre- and post-IPC insertion.

CONCLUSIONS

LENT score seems to be an easy and attainable tool, capable of predicting the survival of the patients with MPE quite accurately. It can be helpful in palliating the symptoms of patients with advanced malignancies by modifying the treatment strategies.

Malignant pleural effusion survival prognostication in an Asian population

BACKGROUND AND OBJECTIVE

LENT and PROMISE scores prognosticate survival in patients with MPE. Prognostication guides the selection of interventions and management. However, the predictive value of these scores and their refinements (modified-LENT) in Asians remain unclear. We aim to evaluate the performance of LENT, modified-LENT and clinical PROMISE scores; identify predictors of survival; and develop an alternative prognostication tool should current scores lack accuracy.

METHODS

Retrospective medical record review of an Asian pleuroscopy database from 2011 to 2018 of patients with MPE was conducted. The prognostic capability of current available scores were evaluated using C-statistics. Demographic and clinical variables as predictors of survival were assessed, and an alternative model was developed using logistic regression.

RESULTS

In 130 patients, the C-statistics for modified-LENT was not significantly different from LENT (0.59 (95% CI: 0.52-0.67) vs 0.56 (95% CI: 0.49-0.63); P = 0.403). In 57 patients, the PROMISE C-statistics was 0.72 (95% CI: 0.53-0.91). In our alternative prognostication model (n = 147), Sex, Eastern Cooperative Oncology Group status, Leukocyte count, EGFR mutation, Chemotherapy and primary Tumour type (SELECT) were predictors of 90-day mortality (C-statistic = 0.87 (95% CI: 0.79-0.95)). SELECT sensitivity, specificity, positive and negative predictive values, and positive and negative likelihood ratios using a predicted probability of 90-day mortality cut-off point of 10% were 0.91, 0.68, 0.34, 0.98, 2.83 and 0.13, respectively.

CONCLUSION

The LENT, modified-LENT and PROMISE scores have poor accuracy of survival prognostication in Asian patients with MPE undergoing pleuroscopy. The proposed SELECT prognostication model is accurate at identifying patients with high probability of survival at 90 days.