Development and validation of nomograms to predict the survival probability and occurrence of a second primary malignancy of male breast cancer patients: a population-based analysis

Abstract

Background: Male breast cancer (MBC) is rare, which has restricted prospective research among MBC patients. With effective treatments, the prognosis of MBC patients has improved and developing a second primary malignancy (SPM) has become a life-threatening event for MBC survivors. However, few studies have focused on the prognosis of MBC patients and looked into the SPM issue in MBC survivors.

Method: We reviewed MBC patients diagnosed between 1990 and 2016 from the latest Surveillance, Epidemiology, and End Results (SEER) Plus database. Competing risk models and nomograms were conducted for predicting the risk of cancer-specific death and SPM occurrence. C-indexes, calibration curves, ROC curves, and decision curve analysis (DCA) curves were applied for validation.

Result: A total of 1,843 MBC patients with complete information were finally enrolled and 60 (3.26%) had developed an SPM. Prostate cancer (40%) was the most common SPM. The median OS of all the enrolled patients was 102.41 months, while the median latency from the initial MBC diagnosis to the subsequent diagnosis of SPM was 67.2 months. The patients who suffered from an SPM shared a longer OS than those patients with only one MBC (p = 0.027). The patients were randomly divided into the development cohort and the validation cohort (at a ratio of 7:3). The Fine and Gray competing risk model was used to identify the risk factors. Two nomograms were constructed and validated to predict the 5-year, 8-year, and 10-year survival probability of MBC patients, both of which had good performance in the C-index, ROC curves, calibration plots, and DCA curves, showing the ideal discrimination capability and predictive value clinically. Furthermore, we, for the first time, constructed a nomogram based on the competing risk model to predict the 5-year, 8-year, and 10-year probability of developing an SPM in MBC survivors, which also showed good discrimination, calibration, and clinical effectiveness.

Conclusion: We, for the first time, included treatment information and clinical parameters to construct a nomogram to predict not only the survival probability of MBC patients but also the probability of developing an SPM in MBC survivors, which were helpful in individual risk estimation, patient follow-up, and counseling in MBC patients.

Keywords: male breast cancer; nomogram; prognosis; second primary malignancy; survival probability.

Figure 1

The detailed distribution of the SPMs among MBC survivors. Prostate cancer represented 24 (40%) of all SPMs, followed by lung and bronchus at 6 (10.0%), melanoma of the skin at 5 (8.3%), the secondary breast cancer at 4 (6.7%), liver at 3 (5.0%), urinary bladder at 3 (5.0%), kidney and renal pelvis at 2 (3.3%), NHL at 2 (3.3%), pancreas at 2 (3.3%), rectum at 2 (3.3%), and stomach at 2 (3.3%).

Figure 2

The survival analyses. (A) There was no significant difference in OS between the development and validation cohort (p = 0.83). (B) Those who developed an SPM have a significantly longer OS in MBC survivors (p = 0.027).

Figure 3

(A) The nomogram model 1 to predict the 5-year, 8-year, and 10-year survival probability of MBC patients based on the Fine and Gray method. (B) The nomogram model 2 to predict the same survival probability of MBC patients based on the multivariate Cox regression.

Figure 4

(A) Both model 1 and model 2 showed better predictive value than TMN stage in the ROC analyses. (B) The calibration curve of model 1. (C) The calibration curve of model 2. (D) The DCA of model 1 in the development cohort. (E) The DCA of model 1 in the validation cohort. (F) The DCA of model 2 in the development cohort. (G) The DCA of model 2 in the validation cohort.

Figure 5

(A) The survival curves between different risk groups in the development cohort in model 1. (B) The survival curves between different risk groups in the validation cohort in model 1. (C) The survival curves between different risk groups in the development cohort in model 2. (D) The survival curves between different risk groups in the validation cohort in model 2.

Figure 6

(A) The nomogram of model 3 for predicting the 10-year probability of MBC survivors who suffer from an SPM. (B) The ROC curve of model 3. (C) The calibration curve of model 3. (D) The DCA of model 3 in the development cohort. (E) The DCA of model 3 in the validation cohort.