The Future Is in the Details, and a Farewell

I am typically more worried about a solid round mass than an oval circumscribed mass seen in any modality. While BI-RADS (1) assessment category 3 has relatively well-defined indications for mammography and US, the indications are less clear for MRI. In fact, the BI-RADS Atlas states, “The use of category 3 assessment at MRI remains intuitive for radiologists who lack extensive (audited) personal experience with any specific type of lesion” (1).

In this issue of the Journal of Breast Imaging, Myers et al (2) provide a retrospective review of 165 circumscribed masses detected on MRI in 158 women, finding an overall malignancy rate of 5.5% (9/165). However, circumscribed masses with round shape had a higher risk of malignancy (7/37, 18.9%) compared with oval shape (2/128, 1.7%) (P < 0.001). Of the two cancers with oval shape, both had washout kinetics such that no oval masses with persistent, plateau, or below the color kinetic threshold were malignant (0/92, 0%). An accompanying editorial by Grimm (3) recognizes the major contribution of this detailed and larger data set to the literature but cautions that we may not be able to apply the same BI-RADS 3 criteria from mammography and US to MRI. Given that the indications for MRI are most commonly evaluating extent of a known breast cancer or screening of women at high risk, following breast lesions is often not clinically appropriate or personally acceptable to the patient. In addition, breast MRI has only fair to moderate reproducibility of descriptors in studies of both intra- and inter-reader variability.

A challenge for our literature is that studies of masses with round shape often do not report the mass margins, or data for round and oval shape are pooled. A recent scientific review in JBI by Nguyen et al (4) notes most studies of BI-RADS 3 for breast MRI pool round and oval shape together but also notes three studies that separate them, finding a higher risk of malignancy for round masses with circumscribed margins compared with oval (5–7). The contribution of this article by Myers et al (2) similarly builds our knowledge with this deeper level of data for masses detected on MRI.

The need to keep developing scientific literature on this topic becomes even more important as we broaden the use of MRI to an abbreviated format for the screening of women with dense breast tissue. Marshall et al (8) describe the results of their first two years after implementing an abbreviated breast MRI program, detecting 16 cancers in 1338 patients (12.0/1000), with one cancer detected following a BI-RADS 3 assessment (1/121, 0.8%).

A study regarding management of palpable circumscribed masses on US by Mahboubi-Fooladi et al (9) reports results of a 20-question survey of members of the Iranian Society of Radiology. Interestingly, radiologists with more than five years of experience and those who had completed a fellowship in breast imaging selected biopsy for palpable circumscribed masses for average-risk women more frequently than their less experienced and non–fellowship-trained counterparts (9). The authors postulate that application of BI-RADS 3 may vary by the experience and education of radiologists as well as cultural norms given that a national screening program has not yet been established in Iran.

Our Scientific Review on the integration of artificial intelligence (AI) into clinical breast imaging (10) further builds on our excellent base of JBI articles on AI (11–18). This review by Ozcan et al (10) discusses the challenges of data privacy, data security, transparency, and potential bias, along with the lack of annotated data sets to establish ground truth. These limitations of AI are demonstrated in a retrospective study of the initial one-year experience with an AI detection tool by Letter et al (19), which finds a nonsignificant trend in improvement in cancer detection rate with no change in abnormal interpretation rate. These results may be disappointing, although the authors note that this tool was used in a subspecialized academic breast program and their findings may be limited in generalizability (19). Additional similar studies on the efficacy of AI detection tools will be useful.

A research article by Ng et al (20) from the UK explores the use of AI as a second reader of screening mammograms compared with human double reading. The study finds that AI as a supporting reader had superior or noninferior results for all screening metrics. This approach may reduce cost and time to final results in screening programs that currently use human double reading.

Two research articles pertain to the COVID-19 pandemic. In the first, Dodelzon et al (21) surveyed breast radiologist members of the Society of Breast Imaging (SBI) and compared results to a survey conducted early in the pandemic. They found that half of the respondents reported two or more psychological distress symptoms, only declining from about 60% in the initial survey. Ongoing distress was particularly related to ongoing pandemic-specific childcare needs as well as financial challenges. A second article by Sivanushanthan et al (22) finds that while both screening mammography and screening MRI rebounded quickly after the initial shutdown during the pandemic, the rebound was not sustained for screening breast MRI. The authors conclude that specific interventions may be needed to promote a return to screening MRI for high-risk women. A Clinical Practice article in this issue of JBI by Gong et al (23) describes the use of patient navigators to address disparities in screening as well as COVID-19 pandemic–specific barriers that continue for many of our patients.

Our remaining review articles include a wealth of information. Our Science of Screening article is a high-level evaluation of molecular breast imaging in the screening setting by Smith et al (24). Kapoor et al (25) provide a Radiologic–Pathologic Correlation review of both primary and secondary etiologies of angiosarcoma with beautiful images and discussion. Our Educational Review is a very informative discussion and illustration of the new 2022 Food and Drug Administration recommendations for breast implant evaluation by Le-Petross et al (26). A review of the care of male patients with breast cancer by Constantinou et al (27) broadens our perspectives. Lawson et al (28) have given us a very practical discussion on obtaining breast imaging research funding that will likely serve as a resource for academic breast radiologists for many years to come.

Farewell

At an SBI leadership strategic retreat in 2017, very strong support was voiced to found a society-based journal that would improve the visibility and discussion of breast imaging research. As the then-head of the SBI Scientific Advisory Committee, I took on the groundwork of starting a journal while consistently messaging that being a journal editor was not a personal goal. At the 2018 SBI Annual Meeting in Las Vegas, we interviewed publishers and made the decision to work with Oxford University Press. At that meeting, it became clear that we could not move forward with the journal without an editor. At that moment and even to this day, I felt so strongly committed to the science of breast imaging that I agreed to be the founding editor. Jay Parikh, Martha Mainiero, and Ana Lourenco were our first Associate Editors, and we sweated together to start the journal. Our first issue published just one year later, in time to be available at the 2019 SBI Annual Meeting in Hollywood, Florida. Our expanding list of Associate Editors and our Editorial Board members, reviewers, and authors have invested countless hours and considerable energy to build the JBI to become a reliable source for breast imaging research and review articles. Thank you to those of you who purposely chose to submit your work to JBI in order to build us up. I also personally want to thank those of you who said “yes” when we asked for a rapid turnaround on a paper or a review. And a special thank you goes to my long-time friend, Jay Baker, who filled in when needed and has given me excellent feedback on my editorials.

I have loved serving as editor. Prior to JBI, the breast imaging community had so few journals available to publish science. Our research is now blossoming. In addition, I love teaching and mentoring, and the JBI has given me so many opportunities to help our authors, reviewers, and journal leaders grow and expand their skills. In turn, I have learned so very much from all of you. It takes a village. The journal now moves on with our new fantastic Editor-in-Chief, Wendy DeMartini, who will bring the journal forward to new heights.

It has been my deepest honor to serve you and our science.

References