To our knowledge, this is the first study examining surveillance strategies and local recurrence detection methods in patients undergoing BCS for ILC. There are several notable findings from the analysis. First, although mammography is known to have poor sensitivity for patients with ILC, only a minority of patients had supplemental breast MRI as part of their surveillance strategy. The only factor associated with undergoing breast MRI was younger age at primary ILC diagnosis; while breast density was not available in this cohort, the association between younger age and higher breast density raises the possibility that density impacted surveillance imaging recommendations. Indeed, current guidelines from the American College of Radiology suggest adding breast MRI for surveillance in those with a prior history of breast cancer and dense breasts.16 Additionally, breast MRI may be utilized for patients who were previously diagnosed with mammographically occult breast cancer. ILC has a higher rate of mammographically occult cancers, as well as a higher rate of understaging on mammography compared with IDC.10,11 Because this study lacks an IDC cohort, we cannot compare rates of MRI surveillance by histologic subtype. However, we hypothesize that breast MRI may have been utilized more often in patients with mammographically occult ILC, or in cases with large discrepancies between clinical tumor size and pathologic tumor size.

Interestingly, among those having mammographic surveillance, most local recurrences presented as interval cancers (61.9%). This finding suggests that although ILC is thought to be less likely to form a palpable mass, physical examination (either self-examination or clinical breast examination) did result in the detection of most recurrences in this subset of cases. It is important to note that a lack of adherence to recommended screening guidelines may bias the results against the utility of mammography, as not all patients underwent mammograms yearly. This is reflected in the length of time between the last mammogram and diagnosis of recurrence in several of the interval cancer cases. However, in the group that had supplemental surveillance with MRI, the majority of recurrences were detected by MRI and none were detected by mammography. Together, these findings suggest that routine mammographic screening may have limited utility for surveillance in patients treated for ILC. This finding is somewhat consistent with data showing that in the setting of screening patients without prior breast cancer, mammography has lower sensitivity for detecting ILC.22 Because of the poor sensitivity for detection, those with ILC are diagnosed at later stages than those with IDC.

Although this study is limited by a small sample size, the finding that most local recurrences in the breast MRI group were in fact detected by MRI is consistent with prior literature showing a higher cancer detection rate with breast MRI compared with mammogram, although prior studies did not specifically study ILC.12,21,22 We were unable to determine biopsy rates and false positive rates because we only included patients who were diagnosed with recurrence. Future analysis of the larger cohort without recurrence may help determine test performance in this study population. Indeed, studies reporting higher rates of biopsies with no impact on overall survival make some clinicians question the benefit of supplemental screening with breast MRI.22,23 However, others have argued that because earlier detection of recurrence could lead to better clinical outcomes, the downsides of breast MRI may be tolerable to many patients.13,24,25 While it is difficult to argue for surveillance breast MRI for all patients with ILC based on the small number of patients undergoing MRI surveillance in our study, the large difference in the interval cancer rate in the mammography and MRI groups, despite the small cohort, suggests that MRI might have particular utility for those with ILC. Indeed, given the lower sensitivity of mammography specifically in ILC, the potential added benefit of breast MRI is reasonably hypothesized to be greater. Of note, in a recent survey study of radiologists, the majority recommended supplemental screening with breast MRI for those with a personal history of ILC.22

The findings in our study suggest that an alternative to conventional mammography is needed for surveillance in ILC patients. One promising solution may be contrast-enhanced mammography, with recent studies showing performance in ILC near that of breast MRI.26 This may be a more feasible imaging tool as availability expands. Larger studies are needed both to validate our findings and to investigate novel imaging tools in those with ILC.

The palpable nature of many of the recurrences is reassuring, as ILC is often thought to be non-palpable. This underscores the utility of physical examination and is also reassuring for patients who undergo mastectomy for ILC, where routine imaging for surveillance is not recommended. More data are needed on the rate of palpable tumors in ILC at diagnosis and differences in tumor biology between primary and locally recurrent tumors.

Strengths of this study include the unique nature of these data with a curated review of individual patient charts to determine the surveillance detection method, which can be very difficult to accurately assess since imaging is always obtained to evaluate clinically reported findings. Careful attention was paid to determine the very first presentation of local recurrence, but the retrospective nature of this study raises the possibility that the recurrence detection method was attributed incorrectly in some cases despite careful review. Additionally, because local recurrence occurs in a minority of patients after BCS for ILC, the overall sample size is small. This reduces the statistical power to identify differences in overall survival by surveillance strategy; although we did not find a difference between groups, there is still a possibility that earlier detection of local recurrence could impact both subsequent treatment and long-term outcomes for these patients.

Overall, we found that for patients who undergo BCS after diagnosis of ILC, supplemental imaging surveillance with breast MRI is more common in younger women and is associated with significantly lower interval cancer rate compared with mammography alone. These findings suggest that surveillance MRI may have utility for those with a prior personal history of ILC, which should be validated in larger studies.