External evaluation of the Briganti nomogram to predict lymph node metastases in intermediate-risk prostate cancer patients

The role and controversy of pelvic lymph node dissection in prostate cancer treatment: a focused review

Pelvic lymph node dissection (PLND) is commonly performed alongside radical prostatectomy. Its primary objective is to determine the lymphatic staging of prostate tumors by removing lymph nodes involved in lymphatic drainage. This aids in guiding subsequent treatment and removing metastatic foci, potentially offering significant therapeutic benefits. Despite varying recommendations from clinical practice guidelines across countries, the actual implementation of PLND is inconsistent, partly due to debates over its therapeutic value. While high-quality evidence supporting the superiority of PLND in oncological outcomes is lacking, its role in increasing surgical time and risk of complications is well-recognized. Despite these concerns, PLND remains the gold standard for lymph node staging in prostate cancer, providing invaluable staging information unattainable by other techniques. This article reviews PLND's scope, guideline perspectives, implementation status, oncologic and non-oncologic outcomes, alternatives, and future research needs.

The predictive value of machine learning and nomograms for lymph node metastasis of prostate cancer: a systematic review and meta-analysis

BACKGROUND

In clinical practice, there are currently a variety of nomograms for predicting lymph node metastasis (LNM) of prostate cancer. At the same time, some scholars have introduced machine learning (ML) into the prediction of LNM of prostate cancer. However, the predictive value of nomograms and ML remains controversial. Based on this situation, this systematic review and meta-analysis was performed to explore the predictive value of various nomograms currently recommended and newly-developed ML models for LNM in prostate cancer patients.

EVIDENCE ACQUISITION

Cochrane, PubMed, Embase, and Web of Science were searched up to November 1, 2022. The risk of bias in the included studies was evaluated using the Prediction model Risk of Bias Assessment Tool (PROBAST). The concordance index (C-index), sensitivity, and specificity were adopted to evaluate the predictive accuracy of the models.

RESULTS

Thirty-one studies (18,803 patients) were included. Seven kinds of nomograms currently recommended, dominated by Briganti nomogram or MSKCC nomogram, were covered in the included studies. For newly-developed ML models, the C-index for LNM prediction in the training set and validation set was 0.846 [95%CI (0.818, 0.873)] and 0.862 [95%CI (0.819-0.905)] respectively. Most ML models in the training set were based on Logistic Regression (LR), which had a sensitivity of 0.78 [95%CI (0.70, 0.85)] and a specificity of 0.85 [95%CI (0.77, 0.90)] in the training set, and a sensitivity of 0.81 [95%CI (0.67, 0.89)] and a specificity of 0.82 [95%CI (0.75, 0.88)] in the validation set. For the recommended nomograms, the C-index in the validation set was 0.745 [95%CI (0.701, 0.790)] for the Briganti nomogram and 0.714 [95%CI (0.662, 0.765)] for the MSKCC nomogram.

CONCLUSION

The predictive accuracy of ML is superior to existing clinically recommended nomograms, and appropriate updates can be conducted to existing nomograms according to special situations.

A Web-Based Prediction Model for Cancer-Specific Survival of Elderly Patients Undergoing Surgery With Prostate Cancer: A Population-Based Study

Objective

Prostate cancer (PC) is the second leading cause of cancer death in men in the United States after lung cancer in global incidence. Elderly male patients over 65 years old account for more than 60% of PC patients, and the impact of surgical treatment on the prognosis of PC patients is controversial. Moreover, there are currently no predictive models that can predict the prognosis of elderly PC patients undergoing surgical treatment. Therefore, we aimed to construct a new nomogram to predict cancer-specific survival (CSS) in elderly PC patients undergoing surgical treatment.

Methods

Data for surgically treated PC patients aged 65 years and older were obtained from the Surveillance, Epidemiology, and End Results (SEER) database. Univariate and multivariate Cox regression models were used to identify independent risk factors for elderly PC patients undergoing surgical treatment. A nomogram of elderly PC patients undergoing surgical treatment was developed based on the multivariate Cox regression model. The consistency index (C-index), the area under the subject operating characteristic curve (AUC), and the calibration curve were used to test the accuracy and discrimination of the predictive model. Decision curve analysis (DCA) was used to examine the potential clinical value of this model.

Results

A total of 44,975 elderly PC patients undergoing surgery in 2010–2018 were randomly assigned to the training set (N = 31705) and validation set (N = 13270). the training set was used for nomogram development and the validation set was used for internal validation. Univariate and multivariate Cox regression model analysis showed that age, marriage, TNM stage, surgical style, chemotherapy, radiotherapy, Gleason score(GS), and prostate-specific antigen(PSA) were independent risk factors for CSS in elderly PC patients undergoing surgical treatment. The C index of the training set and validation indices are 0.911(95%CI: 0.899–0.923) and 0.913(95%CI: 0.893–0.933), respectively, indicating that the nomogram has a good discrimination ability. The AUC and the calibration curves also show good accuracy and discriminability.

Conclusions

To our knowledge, our nomogram is the first predictive model for elderly PC patients undergoing surgical treatment, filling the gap in current predictive models for this PC patient population. Our data comes from the SEER database, which is trustworthy and reliable. Moreover, our model has been internally validated in the validation set using the C-index,AUC and the and the calibration curve, showed that the model have good accuracy and reliability, which can help clinicians and patients make better clinical decision-making. Moreover, the DCA results show that our nomogram has a better potential clinical application value than the TNM staging system.

External Validation and Comparison of Two Nomograms Predicting the Probability of Lymph Node Involvement in Patients subjected to Robot-Assisted Radical Prostatectomy and Concomitant Lymph Node Dissection: A Single Tertiary Center Experience in the MRI-Era

Introduction

To externally validate and directly compare the performance of the Briganti 2012 and Briganti 2019 nomograms as predictors of lymph node invasion (LNI) in a cohort of patients treated with robot-assisted radical prostatectomy (RARP) and extended pelvic lymph node dissection (ePLND).

Materials and Methods

After the exclusion of patients with incomplete biopsy, imaging, or clinical data, 752 patients who underwent RARP and ePLND between December 2014 to August 2021 at our center, were included. Among these patients, 327 (43.5%) had undergone multi-parametric MRI (mpMRI) and mpMRI-targeted biopsy. The preoperative risk of LNI was calculated for all patients using the Briganti 2012 nomogram, while the Briganti 2019 nomogram was used only in patients who had performed mpMRI with the combination of targeted and systematic biopsy. The performances of Briganti 2012 and 2019 models were evaluated using the area under the receiver-operating characteristics curve analysis, calibrations plot, and decision curve analysis.

Results

A median of 13 (IQR 9–18) nodes per patient was removed, and 78 (10.4%) patients had LNI at final pathology. The area under the curves (AUCs) for Briganti 2012 and 2019 were 0.84 and 0.82, respectively. The calibration plots showed a good correlation between the predicted probabilities and the observed proportion of LNI for both models, with a slight tendency to underestimation. The decision curve analysis (DCA) of the two models was similar, with a slightly higher net benefit for Briganti 2012 nomogram. In patients receiving both systematic- and targeted-biopsy, the Briganti 2012 accuracy was 0.85, and no significant difference was found between the AUCs of 2012 and 2019 nomograms (p = 0.296). In the sub-cohort of 518 (68.9%) intermediate-risk PCa patients, the Briganti 2012 nomogram outperforms the 2019 model in terms of accuracy (0.82 vs. 0.77), calibration curve, and net benefit at DCA.

Conclusion

The direct comparison of the two nomograms showed that the most updated nomogram, which included MRI and MRI-targeted biopsy data, was not significantly more accurate than the 2012 model in the prediction of LNI, suggesting a negligible role of mpMRI in the current population.

Development of a Radiomic-Based Model Predicting Lymph Node Involvement in Prostate Cancer Patients

Simple Summary

In patients with prostate cancer, lymph node involvement is a risk factor of relapse. Current guidelines recommend extended lymph node dissection to better stage the disease. However, such a surgical procedure is associated with a higher morbidity than limited lymph node dissection. A better selection of patients is thus essential. Radiomics features are quantitative features automatically extracted from medical imaging. Combining clinical and radiomics features, a machine learning-based model seemed to provide added predictive performance compared to state of the art models regarding the risk prediction of lymph-node involvement in prostate cancer patients.

Abstract

Significant advances in lymph node involvement (LNI) risk modeling in prostate cancer (PCa) have been achieved with the addition of visual interpretation of magnetic resonance imaging (MRI) data, but it is likely that quantitative analysis could further improve prediction models. In this study, we aimed to develop and internally validate a novel LNI risk prediction model based on radiomic features extracted from preoperative multimodal MRI. All patients who underwent a preoperative MRI and radical prostatectomy with extensive lymph node dissection were retrospectively included in a single institution. Patients were randomly divided into the training (60%) and testing (40%) sets. Radiomic features were extracted from the index tumor volumes, delineated on the apparent diffusion coefficient corrected map and the T2 sequences. A ComBat harmonization method was applied to account for inter-site heterogeneity. A prediction model was trained using a neural network approach (Multilayer Perceptron Network, SPSS v24.0©) combining clinical, radiomic and all features. It was then evaluated on the testing set and compared to the current available models using the Receiver Operative Characteristics and the C-Index. Two hundred and eighty patients were included, with a median age of 65.2 y (45.3–79.6), a mean PSA level of 9.5 ng/mL (1.04–63.0) and 79.6% of ISUP ≥ 2 tumors. LNI occurred in 51 patients (18.2%), with a median number of extracted nodes of 15 (10–19). In the testing set, with their respective cutoffs applied, the Partin, Roach, Yale, MSKCC, Briganti 2012 and 2017 models resulted in a C-Index of 0.71, 0.66, 0.55, 0.67, 0.65 and 0.73, respectively, while our proposed combined model resulted in a C-Index of 0.89 in the testing set. Radiomic features extracted from the preoperative MRI scans and combined with clinical features through a neural network seem to provide added predictive performance compared to state of the art models regarding LNI risk prediction in PCa.