Breast Cancer Detection Using a Low-Dose Positron Emission Digital Mammography System

Abstract

Purpose

To investigate the feasibility of low-dose positron emission mammography (PEM) concurrently to MRI to identify breast cancer and determine its local extent.

Materials and Methods

In this research ethics board–approved prospective study, participants newly diagnosed with breast cancer with concurrent breast MRI acquisitions were assigned independently of breast density, tumor size, and histopathologic cancer subtype to undergo low-dose PEM with up to 185 MBq of fluorine 18–labeled fluorodeoxyglucose (¹⁸F-FDG). Two breast radiologists, unaware of the cancer location, reviewed PEM images taken 1 and 4 hours following ¹⁸F-FDG injection. Findings were correlated with histopathologic results. Detection accuracy and participant details were examined using logistic regression and summary statistics, and a comparative analysis assessed the efficacy of PEM and MRI additional lesions detection (ClinicalTrials.gov: NCT03520218).

Results

Twenty-five female participants (median age, 52 years; range, 32–85 years) comprised the cohort. Twenty-four of 25 (96%) cancers (19 invasive cancers and five in situ diseases) were identified with PEM from 100 sets of bilateral images, showcasing comparable performance even after 3 hours of radiotracer uptake. The median invasive cancer size was 31 mm (range, 10–120). Three additional in situ grade 2 lesions were missed at PEM. While not significant, PEM detected fewer false-positive additional lesions compared with MRI (one of six [16%] vs eight of 13 [62%]; P = .14).

Conclusion

This study suggests the feasibility of a low-dose PEM system in helping to detect invasive breast cancer. Though large-scale clinical trials are essential to confirm these preliminary results, this study underscores the potential of this low-dose PEM system as a promising imaging tool in breast cancer diagnosis.

ClinicalTrials.gov registration no. NCT03520218

Keywords: Positron Emission Digital Mammography, Invasive Breast Cancer, Oncology, MRI

Supplemental material is available for this article.

© RSNA, 2024

See also commentary by Barreto and Rapelyea in this issue.

Summary

This study demonstrates the feasibility of an organ-targeted PET system that allows the detection of invasive breast cancer using low doses of fluorine 18–labeled fluorodeoxyglucose.

Key Points

• ■ A pilot study including 25 participants with newly diagnosed breast cancer showed that a low-dose positron emission mammography system enabled detection of 24 of 25 (96%) malignant index lesions (19 invasive cancers and five in situ diseases).

• ■ Although the sample size may constrain our ability to discern significant differences, there was a lower rate of detection of false-positive additional lesions with positron emission mammography than MRI (one of six [16%] vs eight of 13 [62%]; P = .14).

Introduction

Early detection of breast cancer is crucial to ensure a lower rate of advanced disease at diagnosis and, therefore, positively impact overall survival (1). Mammography remains the main imaging modality for detecting breast cancer, as it is the only modality that has proven association with a significant reduction in breast cancer–specific mortality (2).

While screening programs using mammography have been adopted worldwide, the detection of incidental nonpalpable breast lesions remains challenging. Detection is particularly challenging in patients with dense breast tissue, which is a well-known factor that reduces the sensitivity of mammography due to the masking effect of overlying dense fibroglandular tissue, which with a density similar to that of most breast cancers, can obscure the cancer visualization (3). Previous data show 46.9% of the screening population have dense breasts (4). Therefore, continuous advances in imaging modalities and techniques are needed. The role of additional imaging modalities whose diagnostic performance remains unaffected by breast density, such as MRI, contrast-enhanced mammography, and molecular breast imaging (MBI), have become essential to overcoming the dense tissue–related limitation of mammography. There is substantial evidence (5–13) demonstrating that these imaging modalities increase sensitivity compared with mammography because, despite breast density, they can provide information about anatomic changes along with neovascular alterations or metabolic activity of the breast tissue. Thus, these modalities have been considered as potential adjunctive modalities to mammography in this setting.

Regarding MBI, several studies have assessed its use as a supplemental screening modality in individuals with mammographically dense breasts. For example, a prospective trial comparing MBI and two-dimensional mammography in 1585 female participants with dense breasts (10) demonstrated that MBI provided an incremental cancer detection rate of 8.8% for all cancers. Similarly, a retrospective review of 1696 female patients undergoing supplemental screening MBI (11) showed that MBI provided an incremental cancer detection rate of 7.7% for all cancers. Moreover, while positron emission mammography (PEM), a type of MBI technology, and MRI have shown comparable breast-level sensitivity, preliminary studies have shown that PEM has greater specificity to detect breast lesions than MRI (91% vs 86%) (14). Therefore, PEM may lower the false-positive rates associated with MRI, which not only increases patient anxiety but also increases health care costs due to the need for additional imaging, including short-term follow-up or even biopsy (15).

While studies have shown high diagnostic performance of MBI (10,11) as well as its cost-effectiveness when used as a supplemental screening tool for individuals with dense breast tissue (16), MBI has historically been deemed unacceptable as a clinical imaging modality due to the associated high radiation dose delivered not only to the breasts but also to other organs, such as the colon, urinary bladder, and gallbladder. This was mainly attributed to radiation distribution throughout the body, unlike mammography, which targets x-rays directly to the breasts (17,18). Nonetheless, a comprehensive review highlights that the risk of MBI-induced cancer in these other organs is negligible (19). Additionally, newly developed systems using reduced administered radiopharmaceutical doses increased the risk-benefit ratio of MBI and, hence, have advanced MBI as a potential modality to address the limitations of mammography (17,20). However, the current effective dose of MBI with gamma cameras (296 MBq) using technetium 99m sestamibi is approximately 2.4 mSv, which is significantly higher than the dose with mammography (12). Furthermore, dose reduction for MBI technetium 99m sestamibi imaging is highly questionable due to an integral element of gamma camera technology, the collimator, which fundamentally limits the sensitivity of gamma cameras. A new organ-targeted PET system was developed that uses coincidence detection of emitted gamma photons, eliminating the need for collimation and allowing low radiation doses that do not exceed those of mammography (21).

This study investigated the feasibility of low-dose PEM concurrently to MRI to identify breast cancer and determine its local extent.

Materials and Methods

Study Design and Sample

A pilot single-center prospective clinical trial was approved by the University Health Network institutional review board, where the full study protocol is available (research ethics board no. #18–5029). All participants provided written informed consent. The study was also registered at ClinicalTrials.gov (NCT03520218) (22).

From December 2019 to August 2022, individuals with newly diagnosed breast cancer who had both baseline PEM and MRI acquisitions without treatment in between the two scans who provided consent were included in the study. The study sample comprised a convenience series of participants who were assigned independently of their mammographic breast density, tumor size, and histopathologic cancer subtype to undergo PEM at either a 37 MBq, 74 MBq, or 185 MBq dose of fluorine 18–labeled fluorodeoxyglucose (¹⁸F-FDG) using an organ-targeted PET system (Radialis PET Imager; Radialis Medical). Considering the existing MBI modalities with gamma camera devices that use an effective dose of 296 MBq of technetium 99m sestamibi (equivalent to approximately 2.4 mSv [12]), as predefined, to assess the feasibility of this new device in detecting breast cancer with a lower dose as compared with this current standard, we initiated the trial with cases at 185 MBq and subsequently reduced the doses to 74 MBq and 37 MBq. Although Radialis Medical has provided support for the study by supplying the equipment and assisting with its maintenance, it is important to note that Vivianne Freitas, the study's principal investigator and the manuscript's first author, is not an employee or consultant of Radialis Medical. Vivianne Freitas retains full control over the data and information submitted for publication.

While breast MRI is not a standard imaging modality in patients with newly diagnosed breast cancer, as predefined, it was an inclusion requirement to allow for a descriptive comparison with a non–density-dependent modality commonly used in the preoperative setting.

The exclusion criteria were as follows: (a) individuals without breast cancer, (b) those with new breast cancer diagnosed while receiving neoadjuvant chemotherapy during PEM acquisition, (c) those who did not complete the study, and (d) those without previous MRI acquisitions.

Imaging Data Acquisition and Technique

Details of the U.S. Food and Drug Administration–cleared organ-targeted PET system to “image and measure the distribution of injected positron-emitting radiopharmaceuticals in human beings for the purpose of determining various metabolic and physiologic functions within the human body” were described previously (21). Figure 1 shows the Radialis PET Imager.

Figure 1:

Photograph shows Radialis PET Imager.

The eligible participants were injected with ¹⁸F-FDG with either 37 MBq, 74 MBq, or 185 MBq, with the activity chosen independently of the clinical case. Each participant rested for 60 minutes to allow for ¹⁸F-FDG uptake before the first set of PEM images. A subsequent imaging session was performed when the ¹⁸F activity decayed to approximately one-fourth of the initial activity (~2 hours after the first set of images and ~4 hours following injection).

Similar to mammography, the PEM images were obtained for each standard view (craniocaudal and mediolateral oblique) or any additional required mammographic view (single image–acquisition time was ~5 minutes at each position).

Both physician and patient preferences influenced the decision to perform the breast MRI. The MRI protocol we utilized consisted of various sequences: (a) an axial three-dimensional gradient-recalled echo T1-weighted sequence without fat saturation before contrast material administration, (b) a T2-weighted sequence with fat suppression using short inversion time inversion-recovery, and (c) axial three-dimensional gradient-recalled echo T1-weighted sequences both before and after contrast material administration with fat suppression. After administering a standard dose of gadolinium-based contrast material (0.1 mmol/kg) followed by at least a 10-mL saline flush, we conducted five postcontrast acquisitions.

Imaging Interpretation

The PEM images acquired 1 and 4 hours after ¹⁸F-FDG administration were reviewed in consensus by two fellowship-trained breast radiologists (V.F. and A.S.) with more than 20 years of experience in breast imaging, blinded to cancer location. Each breast was assessed separately to better delineate the tumor extent in terms of detecting additional ipsilateral and contralateral disease. A per-lesion–based visual assessment was performed, and the type of uptake lesion (mass vs nonmass uptake lesions) was recorded according to the proposed MBI lexicon by Conners et al (23).

Reference Standard

As prespecified before the start of the study, the ground truth of the lesions was determined by percutaneous-guided histopathologic evaluation. This determination was confirmed either by the final postsurgical pathologic results or, for those lesions not surgically excised, a minimum of 12 months follow-up was also used to confirm the benign nature of the finding. The standard follow-up regimen included an annual clinical examination and breast imaging consisting of either mammography alone or mammography supplemented by MRI.

Statistical Analysis

Participant characteristics were described using summary statistics. Continuous variables were summarized using mean (SD) and median (range), and categorical variables were summarized using count (percentage). Characteristics among different doses were compared using the Kruskal-Wallis test for continuous variables and the χ² test or Fisher exact test for categorical variables. The effect of relevant clinical factors on true detection results versus false results per breast was assessed using a univariable logistic regression model. Given the limited sample size, all participants were included in the univariate analysis, regardless of dose. This analysis was intended for secondary or exploratory outcomes.

Per-dose diagnostic accuracy metrics of PEM in diagnosing the index lesion, including sensitivity, specificity, negative predictive value, positive predictive value, and accuracy of PEM, were calculated with 95% CIs using the binomial method. A descriptive comparative analysis of performance between PEM and MRI in detecting additional lesions was also performed using a two-sided Fisher exact test.

All statistical analyses were conducted using R (version 4.2.2; R Foundation for Statistical Computing), and results were considered significant if the P value was less than .05.

Results

Participant Characteristics

Of 36 participants who were enrolled in the trial, 25 (25 of 36; 69%) met the inclusion criteria and formed the study cohort. Eleven participants were excluded due to: (a) an incomplete study with only a single imaging set acquired (n = 5), (b) dropping out before acquiring the PEM image (n = 1), (c) performance of PEM while undergoing neoadjuvant chemotherapy (n = 3), (e) absence of breast malignancy (n = 1), and (f) lack of previous MRI acquisition (n = 1). Figure 2 summarizes the study workflow.

Figure 2:

Diagram shows the study workflow. ILC = invasive lobular carcinoma, NAC = neoadjuvant chemotherapy, PEM = positron emission mammography.

All participants with newly diagnosed breast cancer included in the study were female individuals (median age, 52 years; range, 32–85 years). Twelve participants (12 of 25; 48%) were considered high-risk for breast cancer due to either genetic mutation or a family history of breast cancer. Most participants were peri- or postmenopausal (14 of 25; 56%), two (two of 14; 14%) of which were undergoing hormone replacement therapy. Twenty participants (20 of 25; 80%) were diagnosed with index invasive disease, including 16 (16 of 20; 80%) with ductal carcinoma, four (four of 20; 20%) with invasive lobular cancer, and five (five of 25; 20%) with index in situ disease. The median invasive cancer size following surgical pathologic analysis of the entire cohort was 32 mm (range, 10–120).

Of 25 participants, 10 (40%) participants received 185 MBq of ¹⁸F-FDG, 10 (40%) received 74 MBq of ¹⁸F-FDG, and five (20%) received 37 MBq of ¹⁸F-FDG. No adverse events related to ¹⁸F-FDG, such as contrast material reaction or any type of complication, were observed during the study. Excluding three participants who had surgery delayed secondary to neoadjuvant therapy after the PEM acquisition, the median time interval between the PEM imaging test and surgery was 20 days (range, 6–52 days).

Participants who received 185 MBq were older than those who received 74 MBq and 37 MBq (median age, 60.5 vs 50.5 vs 45 years; P = .02), and there was also a significant difference in the distribution of the histologic types of breast cancer across the three cohorts, with the in situ disease being the index lesion exclusively in participants who received 74 MBq (P = .02). However, no other clinical and pathologic parameters exhibited differences. The clinical and pathologic parameters of eligible participants stratified by the three different injected doses are summarized in Tables 1 and 2.

Table 1:

Clinical Parameters of Eligible Participants Stratified by Three Different Injected Doses

Table 2:

Pathologic Parameters of Eligible Participants Stratified by Three Different Injected Doses

Detection of Index Lesions

Out of a total of 100 completed bilateral sets of PEM images acquired of each breast across two distinct imaging sessions, comparable diagnostic accuracy was observed despite the variation in time (PEM images acquired 1 and 4 hours after ¹⁸F-FDG administration). With low-dose PEM, 24 of 25 (96%) of known index malignant lesions (19 invasive diseases and five in situ diseases) were detected. However, a single 38-mm lobular index cancer was not detected, representing one of 25 (4%) cases when administered with the lowest dose of ¹⁸F-FDG (37 MBq). In contrast, all 25 index lesions were successfully detected at MRI, resulting in an MRI detection rate of 100% (25 of 25). The median invasive cancer size depicted with PEM was 31 mm (range, 10–100). The masslike lesion type prevailed as the most common lesion in all cohorts, with nonmass lesions observed only in participants who received 74 MBq of ¹⁸F-FDG (P = .008).

Three additional in situ grade 2 malignant lesions were identified in the postsurgical final pathologic analysis, which were not depicted at PEM. On the other hand, two of these extra in situ diseases were not detected at MRI. No clinically relevant factors significantly affected PEM performance.

PEM performance results are summarized in Tables 3 and 4. The per-breast univariable analysis comparing the effect of relevant clinical factors on PEM performance is presented in Tables S1 and S2. We found no evidence of a difference in tumor size between PEM and MRI and the postsurgical pathologic analysis reference standard (Tables S3 and S4).

Table 3:

Summary of Positron Emission Mammography Performance and Accuracy

Table 4:

Per-Dose Positron Emission Mammography Diagnostic Accuracy

Detection of Additional Lesions

A total of six additional lesions were identified with PEM, whereas 13 additional lesions were identified with MRI. All additional lesions noted at PEM were also observed at MRI.

There was a nonsignificant lower rate of detecting false-positive additional lesions with PEM rather than with MRI (one of six [16%] vs eight of 13 [62%]; P = .14). Among the six additional lesions that were detected at PEM, one (16%) false-positive lesion was diagnosed which was subsequently diagnosed as proliferative fibrocystic changes based on the pathologic findings. In contrast, out of the 13 additional lesions detected with MRI, eight (62%) were identified as false positives based on subsequent pathologic examination. For lesions not surgically excised, a minimum of 12 months follow-up was also used to confirm the benign nature of the finding. The confirmed benign nature of these additional lesions detected at MRI included findings such as proliferative fibrocystic changes (n = 2), normal breast tissue (n = 1), intraductal papilloma (n = 1), fibroadenomas (n = 2), hyperplasia of usual type (n = 1), and atypical ductal hyperplasia (n = 1). Figures 3–5 illustrate index cancer lesions and additional lesions detected by PEM compared to mammography and MRI.

Figure 3:

Images obtained in a 50-year-old female patient with a new biopsy-proven malignant lesion in the left breast. (A) Craniocaudal mammogram of the right breast does not show any lesion. (B) The malignant lesion corresponds with a 7.0-cm irregular and spiculated mass on the left craniocaudal mammogram. US-guided core-needle biopsy revealed grade 2 invasive lobular carcinoma. (C) The bilateral positron emission mammographic craniocaudal color image obtained 1 hour after intravenous injection of 185 MBq of fluorine 18–labeled fluorodeoxyglucose (¹⁸F-FDG) shows a mass with intense uptake in the left breast with known cancer and no abnormal uptake in the right breast. Positron emission mammographic craniocaudal images of the left breast obtained (D) 1 hour and (E) 4 hours after intravenous injection of 185 MBq of ¹⁸F-FDG show no substantial visual difference in uptake of the known cancer. (F) Axial contrast-enhanced fat-saturated subtracted T1-weighted MR image with maximum intensity projection reconstruction obtained 90 seconds after intravenous injection of 0.1 mmol of gadolinium-based contrast material per kilogram of body weight also shows the enhancing mass corresponding to known malignancy (arrow) and marked bilateral background parenchymal enhancement, with multiple nonspecific foci of enhancement in the contralateral breast. The patient opted for bilateral mastectomy, which confirmed left-sided malignancy and no malignancy in the contralateral breast.

Figure 5:

Images obtained in a 69-year-old female patient with a new biopsy-proven malignant lesion in the left breast. (A, B) The malignant lesion corresponds with a 2.5-cm irregular and spiculated mass (arrow) and an additional oval and circumscribed mass (double arrows) on left mediolateral oblique and craniocaudal spot compression mammographic views. US-guided core-needle biopsy revealed grade 3 invasive ductal carcinoma (arrow) and fibroadenoma (double arrow). Left positron emission mammographic (PEM) (C) craniocaudal and (D) mediolateral oblique color images obtained 1 hour after intravenous injection of 74 MBq of fluorine 18–labeled fluorodeoxyglucose (¹⁸F-FDG) show a mass (known cancer) with intense uptake and no uptake in the biopsy-proven fibroadenoma. Right PEM (E) craniocaudal and (F) mediolateral oblique color images obtained 1 hour after intravenous injection of 74 MBq of ¹⁸F-FDG do not show any abnormal uptake. Left PEM craniocaudal images obtained (G) 1 hour and (H) 4 hours after intravenous injection of 74 MBq of ¹⁸F-FDG show no substantial visual difference in uptake of known cancer. (I) Axial contrast-enhanced fat-saturated subtracted T1-weighted MR image with maximum intensity projection reconstruction obtained 90 seconds after intravenous injection of 0.1 mmol of gadolinium-based contrast material per kilogram of body weight also shows the left-sided enhancing mass (arrow) corresponding to known malignancy and the biopsy-proven fibroadenoma (double arrows), as well as multiple nonspecific foci of enhancement in the contralateral breast, with one being the most conspicuous (arrow). The patient underwent bilateral lumpectomy. The final pathologic results in the left breast confirmed grade 3 invasive ductal carcinoma and yielded atypical ductal hyperplasia in the right breast.

Figure 4:

Images obtained in an 85-year-old female patient with a new biopsy-proven malignant lesion in the right breast. (A) The malignant lesion corresponds with a 3.0-cm equal-density mildly irregular mass on right craniocaudal mammogram. US-guided core-needle biopsy revealed grade 3 invasive intracystic papillary carcinoma. (B) Right positron emission mammographic (PEM) craniocaudal color image obtained 1 hour after intravenous injection of 74 MBq of fluorine 18–labeled fluorodeoxyglucose (¹⁸F-FDG) shows a 3.2-cm mass (known cancer) with intense uptake. Left PEM (C) mediolateral oblique and (D) craniocaudal color images obtained 1 hour after intravenous injection of 74 MBq of ¹⁸F-FDG do not show any abnormal uptake. Right PEM craniocaudal images obtained (E) 1 hour and (F) 4 hours after intravenous injection of 74 MBq of ¹⁸F-FDG show no substantial visual difference in uptake of the known cancer. (G, H) Axial contrast-enhanced fat-saturated subtracted T1-weighted MR images with maximum intensity projection reconstruction obtained 90 seconds after intravenous injection of 0.1 mmol of gadolinium-based contrast material per kilogram of body weight also show the enhancing mass corresponding to known malignancy and multiple nonspecific foci of enhancement in the contralateral breast, the largest in the left central breast with washout kinetics (arrow in H). Left-sided MRI-guided biopsy confirmed benign intraductal papilloma. The patient underwent a right lumpectomy with sentinel lymph node biopsy. The final pathologic result confirmed 3.2-cm grade 3 invasive intracystic papillary carcinoma with negative sentinel lymph node biopsy.

Discussion

New breast cancer screening strategies to overcome the dense tissue–related limitation of mammography have highlighted the importance of additional imaging modalities whose diagnostic performance remains unaffected by breast density, including MRI, contrast-enhanced mammography, and MBI (5–13). The results of our feasibility study of 25 participants with newly diagnosed breast cancer showed that a PEM system, while historically prohibitive due to the high associated radiation dose, may be a promising imaging modality in this setting, enabling detection of invasive breast cancer with low a ¹⁸F-FDG dose.

The current standard dose of the ¹⁸F-FDG radiotracer for PET imaging is 370 MBq, equal to a whole-body exposure of approximately 6.2–7.1 mSv (24). Although we did not observe significant differences among the various low FDG doses (P = .20), most likely due to the limited sample size underpowering the study, it is noteworthy that we successfully reduced the dose to 37–185 MBq with PEM. This reduction corresponded to a decrease in radiation exposure of 0.62–0.71 mSv to 1.24–1.42 mSv. Notably, the decrement in dose did not hinder any detection of invasive disease, suggesting the plausible feasibility of utilizing lower FDG doses with this PEM device without compromising effectiveness.

In comparison, the average effective dose of natural background radiation in the United States per year is about 3 mSv (25,26). Furthermore, taking into account a mean compressed breast thickness of 5.4 cm (27) and using the International Commission on Radiological Protection conversion factor of 0.12 (28), this low-dose PEM device is comparable to other imaging modalities associated with ionizing radiation, approaching the mean total effective dose of two-view bilateral full-field digital mammography (FFDM), which is about 0.44 mSv (24). The low dose of the PEM device is also comparable with the mean total effective dose of approximately 0.49 mSv used when extra views are needed beyond the standard four FFDM views, which is expected in 20.7% of cases (28). It is also similar to the 0.58 mSv radiation dose of contrast-enhanced mammography, which combines standard FFDM with high- and low-energy images with contrast material (29), and less than the total radiation dose of 0.88 mSv used in digital breast tomosynthesis if a combination of digital breast tomosynthesis and FFDM views are acquired, which is approximately double the dose of FFDM alone (30).

Furthermore, while mainly low-grade in situ disease generated false-negative PEM results, thus impacting PEM sensitivity, Kuhl et al (31,32) have described the clinical uncertainty of this undetected in situ disease at MRI, most likely representing overdiagnosis. These results may be translated to PEM.

In terms of diagnostic performance in detecting additional lesions, although the sample size may constrain our ability to discern significant differences, there was a lower rate of detecting false-positive additional lesions with PEM rather than with MRI (16% vs 62%). It is important to emphasize that the previously documented superior specificity of PEM compared with MRI (14) potentially can remain with low-dose ¹⁸F-FDG administration. These results may represent a potential added benefit of PEM, possibly decreasing both patient anxiety and the cost burden of the health care system related to additional imaging workups, including short-interval follow-ups and/or biopsy. However, to validate our observations, it is crucial to conduct more extensive prospective studies with larger samples (15). Also, although each participant rested for 60 minutes to allow for the ¹⁸F-FDG uptake before the first set of PEM images, the total of 20 minutes to complete the four standard views was a shorter acquisition time than the comparable full MRI protocol, which requires up to 30 minutes (33).

One of the major limitations of our study was the small sample size, which limited the study's statistical power, potentially explaining the lack of significant differences observed among the varied low doses of FDG as well as between the false-positive additional lesion detection rates of PEM and MRI. In addition, the reliance on a convenience sampling method based on those who were willing to participate and the inclusion solely of participants who underwent preoperative MRI might be unrepresentative of the entire population with newly diagnosed breast cancer, thus constraining the generalization of our results. Furthermore, although two breast-trained radiologists reviewed the PEM data, both the overall lesion assessment based on subjective visual parameters and the resolution of divergent interpretation through consensus stand as limitations as they provide an opportunity for those in positions of power to exploit their standing to influence outcomes. Consequently, this may impact the generalizability of our findings. Moreover, although the readers were unaware of the cancer's location, they knew that each participant included in the study had at least one cancer due to the recruitment of only patients with cancer. This knowledge could potentially heighten the probability of detecting cancer. Finally, although our results were descriptively compared with MRI, as another modality unaffected by breast density, comparison with other emerging vascular imaging modalities such as contrast-enhanced mammography is important before considering the clinical applicability of this imaging tool.

In conclusion, our results demonstrate the feasibility of maintaining the detection of invasive breast cancer with a low dose of ¹⁸F-FDG (a dose similar to that of mammography) using PEM system detectors while potentially decreasing false-positive rates associated with MRI. Although large-scale clinical trials are needed to consolidate our preliminary findings, and additional studies are required to identify specific scenarios where its clinical applicability will prove beneficial, this study represents a step forward in making this low-dose PEM system a promising imaging tool in breast cancer diagnosis.