Screening and diagnostic breast MRI: how do they impact surgical treatment? Insights from the MIPA study

Objectives To report mastectomy and reoperation rates in women who had breast MRI for screening (S-MRI subgroup) or diagnostic (D-MRI subgroup) purposes, using multivariable analysis for investigating the role of MRI referral/nonreferral and other covariates in driving surgical outcomes. Methods The MIPA observational study enrolled women aged 18–80 years with newly diagnosed breast cancer destined to have surgery as the primary treatment, in 27 centres worldwide. Mastectomy and reoperation rates were compared using non-parametric tests and multivariable analysis. Results A total of 5828 patients entered analysis, 2763 (47.4%) did not undergo MRI (noMRI subgroup) and 3065 underwent MRI (52.6%); of the latter, 2441/3065 (79.7%) underwent MRI with preoperative intent (P-MRI subgroup), 510/3065 (16.6%) D-MRI, and 114/3065 S-MRI (3.7%). The reoperation rate was 10.5% for S-MRI, 8.2% for D-MRI, and 8.5% for P-MRI, while it was 11.7% for noMRI (p ≤ 0.023 for comparisons with D-MRI and P-MRI). The overall mastectomy rate (first-line mastectomy plus conversions from conserving surgery to mastectomy) was 39.5% for S-MRI, 36.2% for P-MRI, 24.1% for D-MRI, and 18.0% for noMRI. At multivariable analysis, using noMRI as reference, the odds ratios for overall mastectomy were 2.4 (p < 0.001) for S-MRI, 1.0 (p = 0.957) for D-MRI, and 1.9 (p < 0.001) for P-MRI. Conclusions Patients from the D-MRI subgroup had the lowest overall mastectomy rate (24.1%) among MRI subgroups and the lowest reoperation rate (8.2%) together with P-MRI (8.5%). This analysis offers an insight into how the initial indication for MRI affects the subsequent surgical treatment of breast cancer. Key Points • Of 3065 breast MRI examinations, 79.7% were performed with preoperative intent (P-MRI), 16.6% were diagnostic (D-MRI), and 3.7% were screening (S-MRI) examinations. • The D-MRI subgroup had the lowest mastectomy rate (24.1%) among MRI subgroups and the lowest reoperation rate (8.2%) together with P-MRI (8.5%). • The S-MRI subgroup had the highest mastectomy rate (39.5%) which aligns with higher-than-average risk in this subgroup, with a reoperation rate (10.5%) not significantly different to that of all other subgroups. Supplementary Information The online version contains supplementary material available at 10.1007/s00330-023-09600-5.


Imaging
As reported in Table E2, information from conventional imaging was obviously more commonly available in the P-MRI and noMRI subgroups, the P-MRI subgroup having significantly higher rates of patients with cancers with an imaging maximal diameter larger than 20 mm (40% versus 33% at mammography, 31% versus 24% at ultrasonography, respectively) and of patients with multifocal and multicentric cancers (14% versus 9% at mammography, 19% versus 9% at ultrasonography, respectively, adjusted p < 0.001 for all comparisons). Sample paucity and imbalance did not allow to highlight significant pairwise differences in conventional imaging features among the other MRI subgroups, but a relative polarisation of the S-MRI subgroup towards high rates of patients with unifocal (93% versus 87% and 86% at mammography in the D-MRI and P-MRI subgroups, respectively) and small cancers (20% versus 27% and 31% at ultrasonography in the D-MRI and P-MRI subgroups, respectively) can be observed. While no significant difference (p = 0.017) was observed among MRI subgroups in terms of cancer focality at MRI (79% unifocal cancers in the S-MRI subgroup versus 71% and 65% in the D-MRI and P-MRI subgroups, respectively), the rate of cancers larger than 20 mm was markedly lower in the S-MRI (30%) and D-MRI (40%) subgroups (non-significant pairwise comparison with adjusted p = 0.271), compared to 51% in the P-MRI subgroup (adjusted p < 0.001 for both comparisons).

Biopsy and surgical pathology
Comparisons among the subgroups for three indicators from core-needle or vacuum-assisted biopsy (CNB/VAB) and six indicators from surgical pathology are detailed in Table E3 and E4. At biopsy, no significant difference was observed in the rates of occurrence of pure ductal carcinoma in situ (DCIS) among the four subgroups (p = 0.192 at CNB/VAB, p = 0.012 at pathology), the D-MRI and P-MRI subgroups always exhibiting marginally lower rates (15% and 17% at CNB/VAB, respectively, both 14% at surgical pathology) than the S-MRI and noMRI subgroups (21% and 18% at CNB/VAB, 18% and 17% at surgical pathology, respectively).
The rates of the association of DCIS to invasive cancer were instead significantly different among subgroups both at CNB/VAB and surgical pathology (p < 0.001 for both comparisons). At biopsy, this comparison was dominated by the high rate of associated DCIS in the D-MRI subgroup compared to the P-MRI subgroup (37% versus 31%, respectively, adjusted p = 0.003), while subgroup size imbalance hindered the emergence of a significant difference between the 41% rate in the S-MRI subgroup and other subgroups (adjusted p ≥ 0.114 in all comparisons). At surgical pathology, all rates markedly increased and the comparison was dominated by the difference between the 73% rate of DCIS association to invasive cancers in the S-MRI subgroup compared to other subgroups (adjusted p values ≤ 0.010).
The presence of invasive lobular carcinoma-as an exclusive entity or as a component in mixed ductal and lobular cancers-significantly differed among subgroups both at CNB/VAB (p < 0.001) and surgical pathology (p < 0.001). Both comparisons were led by the higher rates of invasive lobular carcinoma occurrence in the D-MRI and P-MRI subgroups (15% and 17% at CNB/VAB, 21% and 22% at surgical pathology, respectively, adjusted p values < 0.001) compared to the S-MRI and noMRI subgroups (10% and 8% at CNB/VAB, 13% and 12% at surgical pathology, respectively).
As already observed at conventional and MRI imaging, the D-MRI and P-MRI subgroups had significantly lower rates of unifocal cancers (77% for both subgroups) compared to the noMRI subgroup (89%, overall and adjusted p values < 0.001), while a similar non-significant trend was also seen for the S-MRI subgroup (83% rate of unifocal cancers, adjusted p values ≥ 0.327).

Multivariable model building
As detailed in the Methods section, variable selection for multivariable logistic regression was performed using stepwise multivariable linear regression (forward selection with p < 0.1 as the threshold for variable inclusion), taking into account a clinically-reasoned pool of variables for each of the four models, each focused on a specific surgical endpoint. Of note, only those information that were effectively available to the multidisciplinary meeting when planning each specific surgical stage were considered for inclusion in the model. More specifically, in the two models investigating first-line mastectomy (Table E5) and bilateral first-line mastectomy (Table E6), variables entering multivariable linear regression were those from demographic, imaging, and CNB/VAB data. In the model investigating reoperation (Table E7), CNB/VAB data were replaced by surgical pathology data from the first surgical procedure. Finally, the model investigating the overall occurrence of mastectomy (Table E8) was built drawing from all aforementioned data categories.