Development and internal validation of a nomogram for predicting pectoralis major fascia invasion in breast cancer patients using preoperative magnetic resonance imaging parameters

Introduction

The incidence of female breast cancer continues to rise annually, making it the most prevalent malignancy threatening women’s health worldwide (1). Modern breast cancer treatment involves a multidisciplinary approach, combining surgery, endocrine therapy, chemotherapy, and radiation therapy, with surgery remaining a central component of care (2,3). In recent years, advancements in surgical techniques and oncologic safety have driven the widespread adoption of immediate breast reconstruction, especially in patients undergoing mastectomy who are ineligible for breast-conserving surgery (BCS) (4). Techniques involving prosthetic reconstruction and acellular dermal matrices have become increasingly common (5-8), with studies showing that preservation of the pectoralis major fascia (PMF) during mastectomy can reduce complications such as capsular contracture, prosthesis displacement, bleeding, and graft rejection (8-10).

Since Halsted first reported on standard radical breast cancer surgery, the PMF has traditionally been considered part of surgical resection. However, some scholars argue that the PMF needs not be removed if the tumor has not yet invaded it (11). The central controversy lies in ensuring oncologic safety when the PMF is preserved, which requires accurate preoperative assessment to determine fascial involvement.

Several studies have investigated PMF preservation and its oncologic outcomes, but they remain limited in design and generalizability. For example, in a Swedish study, patients who underwent PMF-preserving mastectomy had comparable 5-year survival and recurrence rates to those who underwent standard resection (12-14). Similarly, small-sample studies from China with 1–2 years of follow-up found no evidence of local recurrence in patients with preserved PMF (15-17). However, pathological studies have reported invasion rates ranging from 15–26% (18), and these investigations often suffer from limited sample sizes, short follow-up, or a focus on postoperative risk indicators rather than preoperative decision-making tools.

While routine removal of the PMF remains standard practice during mastectomy, emerging reconstructive approaches—particularly those using implants or acellular dermal matrices—have renewed interest in preserving the fascia to improve reconstructive outcomes. Preservation may reduce complications such as capsular contracture, prosthesis malposition, and animation deformity. Nonetheless, due to uncertainty around oncologic safety when invasion is suspected, a method to reliably assess fascial involvement preoperatively remains a clinical necessity

Therefore, this study aims to retrospectively analyze patients’ clinical data, investigate the frequency of tumor invasion of the PMF, explore methods for preoperative assessment of such invasion, and construct a diagnostic tool to predict the risk of tumor invasion of the PMF preoperatively. We present this article in accordance with the TRIPOD reporting checklist (available at https://gs.amegroups.com/article/view/10.21037/gs-2025-186/rc).

Methods

Patients

This was a retrospective, single-center diagnostic study. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the ethics committee of Xuanwu Hospital, Capital Medical University (No. 2024-027-001) and individual consent for this retrospective analysis was waived. We ensured that patient confidentiality was maintained throughout the research process.

We collected data from 235 female breast cancer patients who underwent surgery at our clinic between January 1, 2019, and February 1, 2022. Inclusion criteria were first-time breast cancer treatment and complete clinical data. Of the 235 patients included, 192 underwent modified radical mastectomy and 43 underwent BCS. The primary focus of the study was on the mastectomy group, where the decision to preserve the PMF is most relevant. Exclusion criteria included inflammatory breast cancer, patients receiving neoadjuvant therapy, stage IV breast cancer, specific invasive types such as medullary and mucinous carcinoma, carcinoma in situ, and lack of preoperative breast magnetic resonance imaging (MRI) examination. The gold standard for PMF invasion diagnosis was paraffin-embedded histopathological examination of resected surgical specimens.

Data collection

We recorded clinical data for all patients, including name, age, tumor side, tumor diameter in breast MRI, tumor location, cancer antigen-153 (CA-153) levels, immunohistochemical status, presence of axillary lymph node metastasis, the shortest distance between the tumor and the pectoralis major muscle surface (D1) in breast MRI enhancement delayed phase, the distance from the posterior edge of the tumor-side gland to the pectoralis major muscle surface (D2) in breast MRI lipid suppression phase, D1/D2 ratio, and whether the tumor invaded the PMF in paraffin pathology. The method for measuring D1 and D2 in breast MRI is depicted in Figure 1. Figure 1A shows the delayed phase of breast MRI enhancement, measuring the shortest distance between the tumor and the surface of the pectoralis major muscle (D1 =2.98 cm). Figure 1B shows the breast MRI lipid suppression phase, measuring the distance between the posterior edge of the tumor-side gland and the surface of the pectoralis major muscle (D2 =2.28 cm), resulting in a D1/D2 ratio of 1.31. Statistical analysis was performed using IBM SPSS Statistics 25. The binary logistic regression analysis included variables with P<0.1 in the univariate analysis. Nomogram prediction models were constructed using R 4.0.5, incorporating variables with P<0.1 from the logistic regression analysis. All data were divided into training and validation sets (2:1) for internal validation of the constructed models.

Figure 1 The method for measuring D1 and D2 in breast MRI. (A) The delayed enhancement phase of breast MRI, showing the level of the tumor closest to the surface of the pectoralis major muscle. The red marker indicates the boundary between the tumor and the pectoralis major muscle surface, with the distance between them being D1 =2.98 cm. (B) The breast MRI lipid suppression phase at the nipple level. The red marker indicates the boundary of the breast gland and the boundary of the pectoralis major muscle fascia surface. The tumor is located in the lateral gland, and the distance between them is D2 =2.28 cm. D1/D2 =1.31 was calculated. D1, the shortest distance between tumor and pectoralis major muscle surface in the delayed phase of breast MRI enhancement; D2, the distance between the posterior edge of the tumor-side gland and the surface of the pectoralis major muscle in the lipid-suppressed phase of breast MRI; D1/D2, the ratio of D1 to D2; MRI, magnetic resonance imaging.

Statistical analysis

We used IBM SPSS Statistics 25 and R 4.0.5 for statistical analysis, with a significance level of P<0.1. Postoperative pathological tumor invasion of the PMF was the dependent variable. Measures were tested for normality, and the Chi-squared test was performed if the normal distribution was met. If variances were equal (σ21=σ22), the t-test for two independent sample means was used; otherwise, the t'-test was used. Measures not conforming to a normal distribution were tested using the Mann-Whitney rank sum test. The binary logistic regression analysis included variables with P<0.1 in the univariate analysis. The variables with P<0.1 in the binary logistic regression analysis were used to construct the Nomogram prediction model using R 4.0.5. The study used a random 2:1 ratio to allocate patients into training (n=157) and validation (n=78) cohorts. Due to the retrospective design, formal sample size estimation was not performed. The 95% confidence intervals (CIs) for areas under the curve (AUCs) were calculated using bootstrap resampling.

Results

Basic characteristics

The invasion rate of the PMF was 6.81% in 16 cases out of 235 patients. The mean age of the patients involved was 58 years, without a significant difference between the two groups: tumor invasion (58 years) and no tumor invasion (59 years) (P=0.72). No significant difference was noted in tumor sides (χ²=0.065, P=0.80) or tumor location (χ²=3.776, P=0.43). However, significant differences were noted for:

• Shortest distance from the tumor to the pectoralis major muscle surface (D1) (Z=−4.065, P<0.001);
• Distance from the posterior border of the gland on the tumor side to the pectoralis major muscle surface (D2) (Z=−3.000, P=0.003);
• The D1/D2 ratio (Z=−3.008, P=0.003).

Summary results are represented in Table 1. The tumor diameter on breast MRI did not show any significant difference between groups (Z=−0.459, P=0.64). There was no significant difference noted in the presence of axillary lymph node metastasis (χ²=0.002, P=0.97), vascular cancer embolism (χ²=0.067, P=0.43), elevated CA-153 levels (P=0.62), or immunohistochemical molecular status for hormone receptor (HR) (+/−) (P=0.73) and human epidermal growth factor receptor 2 (Her-2) (+/−) (P=0.65).

Binary logistic regression analysis

Variables included D1, D2, D1/D2. Results showed:

• D1: P=0.10, β=0.938, odds ratio (OR) =2.404, 95% CI: 0.770–8.469;
• D2: P=0.01, β=−2.474, OR =0.091, 95% CI: 0.014–0.585;
• D1/D2: P=0.01, β=−1.496, OR =0.217, 95% CI: 0.071–0.664.

Results are reported in Table 2. The Hosmer-Lemeshow test (χ²=5.499, P=0.78) indicated that the model fits these data well.

Construction of preoperative nomogram prediction model

D1, D2, and D1/D2 were incorporated into the model using R 4.0.5 (Figure 2). The stability of the model evaluated by the Leave-One-Out cross-validation procedure was rather moderate (Kappa =0.21). However, discriminative power was good for the receiver operating characteristic (ROC) curves applied to the training set (AUC =0.866) and the validation set (AUC =0.814) (Figure 3). The model was confirmed to be well-fitting by the Hosmer-Lemeshow test (χ²=4.3233, P=0.78). The calibration curves in Figure 4 show good agreement between predicted and actual probabilities. The decision curve analysis (DCA) indicated clinical utility within the range between the 15% and 30% threshold (Figure 5).

Figure 2 Nomogram model to assess the risk of tumor invasion into the PMF. D1, the shortest distance between tumor and pectoralis major muscle surface in the delayed phase of breast MRI enhancement; D2, the distance between the posterior edge of the tumor-side gland and the surface of the pectoralis major muscle in the lipid-suppressed phase of breast MRI; D1/D2, the ratio of D1 to D2; MRI, magnetic resonance imaging; PMF, pectoralis major fascia.

Figure 3 ROC curve of the prediction model. The y-axis indicates the sensitivity of the model, and the x-axis indicates the specificity of the model. (A) ROC curve of the training set with AUC =0.866. (B) ROC curve of the validation set with AUC =0.814. AUC, area under the curve; CI, confidence interval; OR, odds ratio; ROC, receiver operating characteristic.

Figure 4 Calibration curve of the prediction model. The y-axis represents the actual probability of tumor invasion into the pectoralis major fascia, and the x-axis represents the predicted probability of the model. The calibration curve shows good agreement between the predicted and actual probabilities. (A) The training set calibration curve. (B) The validation set calibration curve.

Figure 5 DCA. The y-axis represents the standardized net benefit, and the x-axis represents the high-risk threshold from 0–0.8. The DCA curve indicates that clinical benefit can be achieved using this assessment tool at a high-risk threshold of 0–0.8. (A) The training set DCA. (B) The validation set DCA. DCA, decision curve analysis.

This is the binary logistic regression model:

P=11+exp[−(0.779+0.877×D1−2.399×D2−1.526×D1D2)][1]

Converting the above equation, it can be expressed in the following linear form:

P=11+exp[−(0.779+0.877×D1−2.399×D2−1.526×D1D2)][2]

Discussion

Breast-enhanced MRI offers unique advantages in assessing the extent of breast tumor distribution and peri-breast anatomy, including the pectoralis major muscle (19,20). Understanding whether the pectoralis fascia is invaded holds significant clinical implications. In oncologic terms, involvement of the fascia may necessitate a broader resection to ensure margin negativity and reduce the risk of local recurrence. From a reconstructive standpoint, retaining the PMF—when oncologically safe—can enhance soft-tissue support for prostheses, improve aesthetic outcomes, and minimize complication rates.

Several studies have examined the issue of PMF preservation in breast cancer. Most of these studies aim to analyze the prognostic impact of surgical preservation of the PMF on patient outcomes through a prospective design combined with long-term follow-up (21,22). The observed endpoints in these studies are mostly overall survival, local recurrence rate, and breast cancer-specific survival.

For example, several studies in Sweden divided patients requiring breast cancer surgery into two groups: those with preservation of the PMF and those with removal of the fascia. These groups were followed for 5 years to compare survival, local recurrence, and chest wall recurrence rates. Follow-up results showed no statistical difference in overall survival, local recurrence rate, or breast cancer-specific survival between the two groups. While preservation of the PMF appeared to increase the risk of chest wall recurrence, the difference between the groups was not statistically significant (23).

The strengths of these studies include the prospective grouping of subjects and the appropriate interventions imposed. However, weaknesses include small sample sizes and the inability to distinguish whether the increased risk of chest wall recurrence is due to preservation of the PMF or tumor heterogeneity. There may be differences in tumor size, aggressiveness, and proliferation ability between patients with preserved and resected PMF. Additionally, some small-sample studies in China followed the prognosis of patients with preserved PMF and found no adverse events, but the small sample size and short follow-up time limit their conclusiveness (24).

Our study acknowledges the shortcomings of previous studies. We utilized preoperative factors that were missing from earlier research and used breast MRI, an imaging technique, to achieve a better preoperative assessment of tumor invasion of the PMF (25,26). The distance from the tumor to the PMF is crucial in imaging. However, due to the prone position nature of breast MRI, the size of the patient’s breast, the level of fat content, and the degree of breast laxity inevitably affect the position of the tumor relative to the PMF. Therefore, we used the thickness of the fat layer between the posterior breast gland and the pectoralis major muscle (i.e., D2 in the final prediction model) as the main influencing factor and used the ratio between the two to minimize the influence of breast volume and laxity.

Although we used the D2 measurement and the D1/D2 ratio to mitigate variability from prone MRI positioning, residual confounding from breast volume, fat content, and tissue laxity could not be entirely eliminated. These anatomical differences may affect the spatial relationship between the tumor and the PMF, potentially impacting measurement accuracy. Additionally, the retrospective nature of this study limits the ability to establish causality. Future studies could incorporate breast volume measurement software or ultrasound-based imaging to enhance positional consistency, although such methods may be costly or difficult to apply in routine clinical settings.

Ultimately, we constructed a prediction model using the relevant risk factors obtained, providing a more convenient tool for clinicians to determine whether to preoperatively preserve the PMF. While the nomogram demonstrated good internal discrimination and calibration, its application in clinical practice warrants cautious interpretation. The model was developed using retrospective data and has not yet been validated prospectively. Further multicenter studies with larger sample sizes are required to confirm its generalizability and real-time predictive value in guiding surgical decisions. However, it is important to note that only 16 patients in our cohort exhibited PMF invasion, representing less than 7% of the sample. This small event rate limits the statistical power and stability of the logistic regression and nomogram model. Although internal validation showed good discrimination, these findings may be prone to overfitting and should be interpreted with caution. Future studies with larger datasets and higher event counts are required for external validation and to strengthen the reliability of the predictive model.

Given that breast MRI is not universally performed in breast cancer patients due to its high cost, there is an inherent selection bias in our retrospective dataset. Patients lacking MRI data were excluded, which may affect the representativeness and external applicability of the nomogram model. This limitation underscores the need for future studies incorporating more broadly available imaging modalities or evaluating cost-effective screening strategies. We did not fundamentally eliminate the effect of breast volume and laxity on the accuracy of the assessment. Future studies could use software that measures breast volume for better assessment, though this is costly. Alternatively, breast ultrasound could be used to eliminate the effect of the prone position examination.

Conclusions

In conclusion, our study shows that the shortest distance between the tumor and the pectoralis major muscle surface in the breast MRI enhancement delayed phase (D1), the distance between the posterior edge of the tumor-side gland and the pectoralis major muscle surface in the breast MRI lipid suppression phase (D2), and the D1/D2 ratio are effective tools for the preoperative differential diagnosis of breast cancer invasion of the PMF. Our prediction model can provide a reliable assessment of the likelihood of tumor invasion of the PMF before surgery.

Acknowledgments

We would like to thank Prof. Wu for his academic support and everyone who supported this research.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://gs.amegroups.com/article/view/10.21037/gs-2025-186/rc

Data Sharing Statement: Available at https://gs.amegroups.com/article/view/10.21037/gs-2025-186/dss

Peer Review File: Available at https://gs.amegroups.com/article/view/10.21037/gs-2025-186/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://gs.amegroups.com/article/view/10.21037/gs-2025-186/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the ethics committee of Xuanwu Hospital, Capital Medical University (No. 2024-027-001) and individual consent for this retrospective analysis was waived.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Łukasiewicz S, Czeczelewski M, Forma A, et al. Breast Cancer-Epidemiology, Risk Factors, Classification, Prognostic Markers, and Current Treatment Strategies-An Updated Review. Cancers (Basel) 2021;13:4287. [Crossref] [PubMed]
2. Ben-Dror J, Shalamov M, Sonnenblick A. The History of Early Breast Cancer Treatment. Genes (Basel) 2022;13:960. [Crossref] [PubMed]
3. Waks AG, Winer EP. Breast Cancer Treatment: A Review. JAMA 2019;321:288-300. [Crossref] [PubMed]
4. Jonczyk MM, Jean J, Graham R, et al. Surgical trends in breast cancer: a rise in novel operative treatment options over a 12 year analysis. Breast Cancer Res Treat 2019;173:267-74. [Crossref] [PubMed]
5. Zaborowski AM, Heeney A, Walsh S, et al. Immediate breast reconstruction. Br J Surg 2023;110:1039-42. [Crossref] [PubMed]
6. O'Shaughnessy K. Evolution and update on current devices for prosthetic breast reconstruction. Gland Surg 2015;4:97-110. [Crossref] [PubMed]
7. Quinn TT, Miller GS, Rostek M, et al. Prosthetic breast reconstruction: indications and update. Gland Surg 2016;5:174-86. [Crossref] [PubMed]
8. Nahabedian MY. Innovations and advancements with prosthetic breast reconstruction. Breast J 2018;24:586-91. [Crossref] [PubMed]
9. Abdelhamid MI, Alkilany MM, Lotfy M. Pectoral fascia preservation during modified radical mastectomy: why and when. The Egyptian Journal of Surgery 2017;36:333-5.
10. Suijker J, Blok YL, de Vries R, et al. Pectoral Fascia Preservation in Oncological Mastectomy to Reduce Complications and Improve Reconstructions: A Systematic Review. Plast Reconstr Surg Glob Open 2020;8:e2700. [Crossref] [PubMed]
11. Vallejo da Silva A, Rodriguez FR, Loures CM, et al. Mastectomy in the era of implant-based reconstruction: should we be removing the pectoralis fascia? Breast 2012;21:779-80. [Crossref] [PubMed]
12. Dalberg K, Johansson H, Signomklao T, et al. A randomised study of axillary drainage and pectoral fascia preservation after mastectomy for breast cancer. Eur J Surg Oncol 2004;30:602-9. [Crossref] [PubMed]
13. Dalberg K, Krawiec K, Sandelin K. Eleven-year follow-up of a randomized study of pectoral fascia preservation after mastectomy for early breast cancer. World J Surg 2010;34:2539-44. [Crossref] [PubMed]
14. Sandelin K, Wickman M, Billgren AM. Oncological outcome after immediate breast reconstruction for invasive breast cancer: a long-term study. Breast 2004;13:210-8. [Crossref] [PubMed]
15. Li J, Sun S. Application analysis of implant coverage with pectoral fascia in immediate implant-based breast reconstruction. J Clin Surg 2019;27:51-4.
16. Zou W, Bai Y, Wang X, et al. Application of pectoralis major fascia in immediate breast 18. reconstruction following mastectomy for breast cancer. Chinese Journal of General Surgery 2017;26:1447-52.
17. Zeng J, Wang Y, Fan P, et al. A preliminary study on retention of the pectoralis major muscle fascia in breast reconstruction with prosthesis at first stage. China Journal of Modern Medicine 2018;28:94-7.
18. Nakao K, Miyata M, Aono T, et al. Cancer cell emboli in the pectoral lymphatics of patients with breast cancer. Jpn J Surg 1989;19:392-7. [Crossref] [PubMed]
19. Thompson JL, Wright GP. The role of breast MRI in newly diagnosed breast cancer: An evidence-based review. Am J Surg 2021;221:525-8. [Crossref] [PubMed]
20. Mann RM, Cho N, Moy L. Breast MRI: State of the Art. Radiology 2019;292:520-36. [Crossref] [PubMed]
21. Paulson D, Bowen ME, Lichtenberg PA. Does brain reserve protect older women from vascular depression? J Gerontol B Psychol Sci Soc Sci 2014;69:157-67. [Crossref] [PubMed]
22. 24. Rutten A, Isgum I, Prokop M. Coronary calcification: effect of small variation of scan starting position on Agatston, volume, and mass scores. Radiology 2008;246:90-8. [Crossref] [PubMed]
23. Mann RM, Kuhl CK, Kinkel K, et al. Breast MRI: guidelines from the European Society of Breast Imaging. Eur Radiol 2008;18:1307-18. [Crossref] [PubMed]
24. Delgado-Alvarado M, Sedano MJ, González-Quintanilla V, et al. Progressive multifocal leukoencephalopathy and idiopathic CD4 lymphocytopenia. J Neurol Sci 2013;327:75-9. [Crossref] [PubMed]
25. Marell DJ, Emond SJ, Kulshrestha A, et al. Analysis of seven-membered lactones by computational NMR methods: proton NMR chemical shift data are more discriminating than carbon. J Org Chem 2014;79:752-8. [Crossref] [PubMed]
26. Sardanelli F, Boetes C, Borisch B, et al. Magnetic resonance imaging of the breast: recommendations from the EUSOMA working group. Eur J Cancer 2010;46:1296-316. [Crossref] [PubMed]