The nipple-areolar complex (NAC) is an integral anatomical region of the breast containing ducts. This anatomical area enables lactation through its unique cellular composition. 123 Approximately 8% of breast cancers originate within the central ducts near the NAC. 34 Diagnosing lesions in this subareolar region are more difficult than in other quadrants, as they may mimic normal nipple anatomy. 1235 Infiltrating ductal carcinoma, ductal carcinoma in situ (DCIS), and Paget's disease can involve the NAC by extending from underlying mammary ducts. 67 Paget's disease involves migrating adenocarcinoma cells, known as Paget cells, from underlying DCIS into the epidermis of the NAC, resulting in a characteristic eczematous presentation during clinical examination. Invasive carcinomas arising in lactiferous ducts can also infiltrate the nipple stroma and dermis. 68910 Examination of mastectomy specimens has revealed higher rates of occult nipple cancer involvement than previously thought, from 8% to 58% in normal-appearing nipples. 4671112 Histopathological analysis demonstrated correlations between nipple cancer infiltration and central tumor location, large size, nodal metastases, lymphovascular invasion, and multicentricity/ multifocality. 1113141516 With broadening selection criteria, meticulous preoperative imaging to assess nipple-areolar and subjacent tumor involvement has become critical for identifying nipple-sparing mastectomy (NSM) candidates to ensure oncologic safety. 6711141718 However, mammograms and ultrasound often underestimate tumor size, while clinical exams and mammograms alone inadequately detect subtle nipple invasion by carcinoma in situ or Paget's disease, necessitating other modalities for staging. 69192021222324 Breast magnetic resonance imaging (MRI) can predict NAC cancer involvement more precisely than mammogram or ultrasound by allowing assessment of NAC shape and contrast enhancement indicative of direct tumor spread. Additionally, MRI enables accurate tumor measurement and localization to inform surgical planning. 6719232526 Numerous studies have shown that tumor proximity to the NAC is a reliable imaging biomarker for cancer infiltration, while nipple enhancement asymmetry on MRI predicts occult involvement. 18[27][28][29][30][31][32][33] Non-mass enhancement (NME) patterns, ipsilateral nipple enhancement, and shorter tumor-nipple distance have also been associated with NAC invasion, serving as MRI indicators of occult involvement in nipple-sparing mastectomy candidates. 1729,31,34,35 This systematic review and meta-analysis investigates MRI which previously has been shown to detect occult NAC malignancy in NSM candidates. By comparing findings across published studies, we aim to determine the most accurate imaging predictors of nipple involvement and optimize MRI to discern subtle neoplastic infiltration better. The importance of MRI in assessing NAC malignancy can have a vital role in both prognostic and diagnostic processes of breast cancers. Due to the high rates of breast cancers, MRI can improve patient selection and surgical planning for safe oncologic outcomes.

This systematic review was conducted according to the 2009 Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA). 36

The electronic databases PubMed, Scopus, Web of Science, and EMBASE were systematically searched to identify all studies on "diagnostic performance of MRI in malignant extension to the nipple" using targeted medical subject headings (MeSH). Two researchers independently searched using customized strategies for each database and assessed all relevant peer-reviewed articles published up to February 2024. This search was conducted using the following keywords: ((Nipple*) OR (Areola*) OR (Nippleareola*)) AND (("Malignant invasion") OR ("tumor enhancement") OR ("tumor invasion") OR ("tumor extension")) AND (("Magnetic Resonance Imaging") OR ("MRI Scan*") OR ("Magnetic Resonance Imaging, Functional")). A review of the references of selected studies was also conducted to identify any additional articles, if the search process had not identified them.

In this systematic review and meta-analysis article, we reviewed all available studies to determine how MRI can be used to assess the invasion of breast malignancies into the NAC and the specificity and sensitivity of MRI, both quantitatively and qualitatively. Initial screening was conducted independently by two authors using titles and abstracts, and two authors chose eligible studies after reading the full texts of the selected articles. To resolve differences of opinion, a third author was consulted. All 5 studies were clinical trials.

Studies examining the utilization of MRI for assessing breast malignancies invading the nipple alveolar complex.

-

A bivariate model was used to obtain summary estimations of sensitivity, specificity, and area under the curve (AUC) data for diagnostic meta-analysis. DerSimonian and Laird random-effects modeling was used to quantify and pool independent diagnostic scores and their variations. 26 We took into account both between-study and within-study variations, both of which contributed to studyweighting. Forest plots were used to show study-specific estimates and 95% confidence intervals (CIs). The I 2 statistic was used to measure statistical heterogeneity, with less than 30% being low, 30% to 60% being moderate, and more than 60% being high. Stata version 15 (Stata Corp LP, College Station, TX, USA) was used for the analyses. P-values that were equal to or less than 0.05 were deemed to be statistically significant.

The authors utilized Egger's funnel plot and Begg's test to analyze the publication bias of the selected papers, with P<0.05 being a significant publication bias. The authors conducted a linear regression analysis for publication bias, including intercept and slope parameters. It was computed using the following equation: yi = a+βxi+ϵi, i = 1… r (r =the number of studies), yi = standardized estimate, xi = precision of studies, ϵi = error terms

The summary of the process of the study selection is depicted in Figure 1 Our search included all articles published before 2024. After applying the selected search term, we found 1870 articles from PubMed, 3490 from Scopus, Web of Science, and Embase. Initially, from the 5360 retrieved reports, 1776 articles were eliminated due to duplication. Based on inclusion and exclusion criteria, 3584 articles were excluded, and the remaining 142 articles were considered for complete text examination. Finally, five articles were selected for our meta-analysis ( Figure 1).

Newcastle-Ottawa scale (NOS) was employed to assess the quality of papers. Based on NOS, each paper was scored a maximum of 9 points based on these criteria: Sample size selection (4 points), comparability (2 points), assessment of the outcomes (3 points). A score of more than 6 was marked as a high-quality study. Studies with moderate quality had scores of 4-6., and the quality of studies scored below 4 was considered to be low (Table 1).

The characteristics of the five studies used for this meta-analysis are demonstrated in Table 1. The publication dates ranged from 2012 to 2020. Two studies were from South Korea, two from Japan, and one from Italy. Overall, the analysis included 653 participants who underwent MRI for the diagnosis of breast cancer. The total number of NAC involvements was 116, with two studies indicating that 24 patients were asymptomatic. Three studies documented 175 individuals in a post-menopausal state, and two out of five studies reported that 17 patients were diagnosed with NAC malignancy at the post-menopausal age.

Alonzo et al. 27 and Hwang et al. 28 disclosed the mean tumor size among patients with NAC involvement, which was 52.6mm and 47.9 ± 23.0mm, respectively. Additionally, they provided data on the mean tumor size in patients without NAC involvement, with values of 51.5 and 39.6, respectively. They also measured the mean distance between the tumor and NAC for positive and negative involvement. In a study by Alonzo et al. 27 , these distances were 7.9 and 23.4, while in a study by Hwang et al. 28 they were 5.3 ± 6.4 and 13.5 ± 13.5, respectively. Yoo et al. also reported these distances to be 29.1 ± 16.7 in the negative NAC involvement group and 13.7 ± 8.7 in patients with NAC involvement.

In our meta-analysis, five articles were deemed eligible to participate in the diagnostic test accuracy (DTA) analysis. The pooled sensitivity of our analysis was found to be 8985% (95% CI: 8177%-9491%), and the heterogeneity within the included studies was approximately low (I2 = 17.5100.00%).

The pooled specificity was also 8982% (95% CI: 7149%-9695%), but our analysis revealed notable heterogeneity in specificity (I2 = 95.87%). ( figure 2.b).

Furthermore, we computed the pooled positive diagnostic likelihood ratio (pDLR) to be 7.814.6 (95% CI: 2.821.3-2116.61, I2 = 9296.66%). This suggests that a positive outcome significantly enhances the probability of accurately determining occult NAC malignancy in nipple-sparing mastectomy candidates. In contrast, the pooled negative diagnostic likelihood ratio (nDLR) was calculated to be 0.12 18 (95% CI: 0.0710-0.2233, I2 = 11.9450.06%) as depicted in Figure 2. Figure 24. Therefore, the diagnostic performance of MRI for detecting occult NAC malignancy is good and efficient. Regrettably, the lack of data made conducting metaregression and subgroup analysis impossible. Figure 3 displays the summary receiver operating curve (SROC), offering a visual representation of the overall diagnostic performance of each study. Based on the bivariate meta-analysis, the summary area under the curve (sAUC) was calculated to be 91% (95% CI: 88%-93%) which is available in Figure 3.

Diagnostic Odd Ratio (DOR) allows us to evaluate the diagnostic accuracy of MRI in NAC malignancy based on the TP, TN, FP, FN. According to our analysis, the DOR of MRI in determining occult NAC malignancy in nipple-sparing mastectomy candidates was 72.21 (95%CI= 15.94 to 326.98, I2=76%) ( Figure 4).

We performed a sensitivity analysis to assess how each study might impact the pooled DOR's result. In sensitivity analysis, it is important to assess the weight of each study and find the outlier studies to prevent high bias. As depicted in Figure 5, the overall pooled DOR was not significantly changed by excluding any included study. The results also indicated that each study did not influence the effect size notably.

To evaluate the potential impact of publication bias on our results, we conducted Deek's funnel plot asymmetry test. The findings from this analysis indicated no substantial publication bias across the diagnostic accuracy indices reported in each of the included studies. (Figure 6)

Five of the research articles featured in the section on meta-analysis MRI were determined to have an 89% sensitivity (95% CI: 81%-94%) and 89% specificity (95% CI: 71%-96%) in detecting occult NAC cancer in our investigation.

As discussed earlier, the rate of unanticipated tumor involvement of the clinically normal nipple ranges from 8% to 58%, according to the analysis of mastectomy tissues. Therefore, estimating the NAC tumour involvement before surgery is critical in determining which individuals would benefit from an NSM. 172930 MRI and mammography have been shown to predict nipple cancer involvement in several investigations. 31 In a study by Byon, Hwang et al. 2023 32 , four diagnostic MRI characteristics-NAC continuity, unilateral NAC enhancement, non-mass enhancement type (refers to lesions accompanied by NME or solely composed of NME, excluding pure mass lesions.), and mass size (greater than 20 mm)-were assessed concerning the presence or absence of NAC involvement. The estimated sensitivity and specificity for NAC involvement using a bivariate random-effects model were as follows: continuity to the NAC (Sensitivity: 71%, Specificity: 94%), unilateral NAC enhancement (Sensitivity: 58%, Specificity: 97%), NME type (Sensitivity: 55%, Specificity: 83%), and mass size > 20 mm (Sensitivity: 88%, Specificity: 58%). They concluded that unilateral NAC enhancement and continuity to the NAC could aid in predicting occult NAC involvement in breast cancer. The outcomes of our investigation are concordant with these observations, indicating a consistent alignment between the findings of our study and the results above. They also concluded that a reasonable cutoff value should be considered to obtain the required diagnostic performance with TND.

In the study of Berger, Luparia et al. 2017 33 , a comparison was made between MRI and galactography for the detection of any lesion in patients with pathologic nipple discharge. However, the evaluation of MRI's diagnostic performance in cancer detection was limited to this modality. In this study, the bivariate showed that the pooled sensitivity for MRI and galactography was 92% and 69%, repectively.

For MRI, the pooled specificity was 76%, while for galactography, it was 39%. In alignment with our study results, they similarly concluded that MRI holds more excellent diagnostic value compared to galactography in any lesion, particularly in cases of isolated breast cancer within the same patient pool. In the detection and diagnosis of breast cancers, alternative modalities including ductoscopy and mammography may also pose competition to MRI.

Another meta-analysis study was conducted by Filipe, Patuleia et al. in 2021. 34 Regarding the diagnosis and treatment of patients with pathological nipple discharge, ductoscopy was compared to MRI. Ductoscopy demonstrated a sensitivity of 44% and a specificity of 98% for breast cancer detection. In comparison, MRI exhibited a sensitivity of 76% and a specificity of 84% for the same purpose. The outcome demonstrated that MRI has greater sensitivity, whereas ductoscopy had considerably greater specificity. In determining the presence of malignancy in individuals exhibiting pathological nipple discharge, their investigation demonstrated that ductoscopy offers a markedly superior diagnosis accuracy in this particular group of patients. Ductoscopy is a more cost-effective and efficient diagnostic method than MRI for identifying patients who need surgery to rule out cancer, which is inconsistent with the findings of our investigations.

In another study, Madsen, Mosebo et al. in 2021 46 compared the diagnostic accuracy of mammography and MRI in identifying the underlying malignancy in a more specific population (Paget's disease) by analyzing the available data. The researchers concluded that incorporating MRI alongside mammography resulted in a considerably higher Mammography exhibited a sensitivity of 39%, while MRI demonstrated a higher sensitivity of 68%, with a significant difference between the two (P=0.0025). Both mammography and MRI had a specificity of 100%, with the specificity range for MRI varying from 29% to 100%. While these results are consistent with our study, it is essential to note that the sample size was small and the population under investigation was specific.

Alaref, Hassan et al. in 2021 analyzed the utility of breast MRI as a screening tool for high-risk women and compared its significance with traditional methods, such as mammography and ultrasound in the early detection of invasive breast cancer. They confirmed that due to its high sensitivity and specificity, breast MRI is superior to mammogram and ultrasound in detecting early breast cancer. On the other hand, its utilization should be avoided when not necessary. Likewise, the findings of this systematic review study are in conformity with the findings of our own standard study about the superiority of MRI in detecting breast cancers. Furthermore, they advocated that developments in breast MRI sequences and techniques, such as the application of diffusion weight sequences, the use of high magnetic fields (up to 7 T), spatial resolution improvements, and spectroscopy, are encouraging instruments that will contribute to MRI's excellence.

Another meta-analysis and systematic review study was conducted by Eisen, Fletcher et al. in 2023. 35 The objective was to provide a comprehensive overview of how the supplementary data regarding the extent of the disease acquired through preoperative breast MRI impacts surgical management, survival, recurrence rates, re-excision, and early detection of bilateral cancer in newly diagnosed breast cancer patients. They reached the conclusion that pre-operative breast MRI significantly reduces the need for reoperations (including re-excisions and conversion to mastectomy) and recurrence. Their findings were consistent with our own; nevertheless, the RCT trial data they utilized were inadequate and posed a significant risk of bias. However, missing the involvement of the nipple alveolar complex and the MRI parameters in predicting the outcome demonstrated a notable difference when compared to our investigation.

This study offers some advantages. First, the samples were screened specifically for malignant invasion into the NAC. Moreover the MRI characteristics were thoroughly examined and compared.

Our analysis had some limitations. Firstly, conducting comparisons was challenging due to the small sample sizes and limited number of studies included. The second limitation was that the majority of articles in this field did not evaluate nipple alveolar complex invasion specifically, so there was a paucity of data regarding NAC malignancy. Thirdly, certain articles did not highlight the specific modality (MRI) that we were evaluating in our research, and others concentrated on the application of the imaging modality for therapeutic purposes rather than diagnosis. Finally, it was not possible to assess the impact of variability in breast MRI sequences and protocols. In Alonzo study 27 , we used data of <5 mm cut-off for patients undergoing MRI (postmenopausal and premenopausal women) and for Hwang 28 study, we used data from strands to nipple in delay phase.

Overall, the included studies showed that MRI had a high diagnostic accuracy strength for identifying the NAC involvement in breast cancer. With a 85% sensitivity of MRI in NAC diagnosis, MRI could be suggested as a beneficial and efficient radiologic tool for diagnosing NAC malignancy. These findings can be used in detection settings to contribute to better diagnosis and prevention of breast cancer.

The Iranian National Committee for Medical Ethics has approved this study.

None.