Comparative diagnostic efficacy of abbreviated and full protocol breast MRI: a systematic review and a meta-analysis

Abstract

Objectives

This meta-analysis compares the efficacy, limitations, and clinical implications of abbreviated breast MRI (AB-MRI) and full protocol MRI (FP-MRI), focusing on diagnostic accuracy across diverse populations. It extends previous analyses by including studies conducted after 2019 in both screening and diagnostic contexts.

Methods

We conducted a systematic review (November 2019 to December 2022), using a bivariate model to calculate summary estimates of sensitivity and specificity. Random effect models were applied for summary area under the curve (AUC), and probability distributions for negative and positive predictive values were obtained. Subgroup analyses explored differences in sensitivity, specificity, and AUC between AB-MRI and FP-MRI.

Results

From 11 eligible studies (1 prospective, 10 retrospective), statistical analysis revealed a significant difference in sensitivity between FP-MRI (95%) and AB-MRI (86%, P = .005), with no significant difference in specificity (P = .50). AB-MRI’s shorter acquisition time suggests potential for higher patient throughput, but challenges remain in detecting small lesions and nonmass enhancements. Some studies recommend additional sequences, like diffusion-weighted imaging, to improve diagnostic performance.

Conclusions

While FP-MRI remains the gold standard in breast cancer detection, AB-MRI offers a quicker alternative, especially in high-risk screening. However, its lower sensitivity limits its use as a standalone diagnostic tool. Future research should optimize AB-MRI protocols and consider patient-specific factors to enhance breast cancer screening and diagnostic strategies.

Advances in knowledge

This meta-analysis expands understanding of AB-MRI's role in breast cancer detection, highlighting its benefits and limitations compared to FP-MRI, particularly in terms of sensitivity and screening efficiency.

Introduction

Globally, breast cancer ranks as the most prevalent cancer among women.¹

Breast MRI stands out as the most effective technique for the early detection of breast cancer and is particularly advised for individuals at elevated risk. There is an ongoing discussion about extending its use as a screening method for women with intermediate risk and dense breast tissue, following evidence from recent research indicating a higher rate of cancer detection in this group.^2,3

As breast MRI begins to be implemented more broadly for screening, particularly in specific subgroups, there is a pressing need to enhance patient throughput without compromising its critical diagnostic accuracy, especially its sensitivity. Since when, 2014, Kuhl et al⁴ introduced the concept of MRI abbreviated protocol (AB-MRI), the interest in this topic has grown significantly, with several studies comparing the AB-MRI versus the full protocol of breast MRI (FP-MRI).

Over the past decade, the trend has been to streamline breast MRI protocols by omitting noncritical sequences, thereby reducing the scan duration to as little as 3 min per examination. This approach could potentially address the challenge of increasing patient throughput. However, the actual impact of shortened scan times on patient throughput remains underresearched.

Moreover, different versions of the AB-MRI have been proposed, and there is no consensus yet for what sequences should be included or excluded from the abbreviated protocol to achieve an optimal diagnostic performance.

The AB-MRI may be useful in specific clinical scenarios, as surveillance imaging for high-risk patients or follow-up evaluations of known lesions, where the need for a comprehensive assessment is less critical. Nevertheless, to date, only 2 meta-analyses^5,6 have been performed that systematically compare the diagnostic performance of AB-MRI with the FP-MRI in studies published before August 2019 and November 2019, respectively.

Our meta-analysis aims to explore and compare the efficacy, limitations, and potential clinical implications of the AB-MRI versus the FP-MRI.

Unlike the meta-analysis performed by Geach et al,⁶ we included studies performed after 2019 not only in the screening setting for high-risk patients but also in the diagnostic one, considering either the general population, high-risk population, or patients with a positive history of breast cancer. Furthermore, despite being aware that the performances of the different studies also vary according to the chosen population, we did not split the selected studies into subgroups as in the meta-analysis performed by Baxter et al,⁵ as our main focus was the overall diagnostic accuracy of the AB-MRI.

Methods

Literature research and study selection

A comprehensive literature search was conducted across multiple databases, including PubMed, to ensure a broad coverage of relevant studies. The search incorporated a combination of terms related to breast cancer and MRI, utilizing both MeSH terms and free text words, as follows: [(“Breast”[Mesh] OR “mammary”[tiab] OR “neoplasm”[tiab] OR “cancer”[tiab] OR “carcinoma”[tiab]) AND (“Magnetic Resonance”[Mesh] OR “MRI”[tiab] OR “abbreviated”[tiab] OR “fast”[tiab] OR “protocol”[tiab])]. The search was conducted from 01 November 2019 to 12 December 2022. A manual search of reference lists from included studies was also undertaken to identify additional relevant publications.

Eligibility criteria

Studies were selected in the systematic review and meta-analysis if they fulfilled the following inclusion criteria:

1. Studies performed in the screening or diagnostic setting.

2. Studies that investigated the diagnostic accuracy of AB-MRI protocol compared to the FP-MRI in either the general population, high-risk population, or patients with a positive history of breast cancer.

The exclusion criteria were:

1. Studies that did not report the sequences of the AB-MRI protocols.

2. Studies that compared the AB-MRI and the FP-MRI for purposes other than their diagnostic accuracy, ie, investigating the efficacy of AB-MRI in neoadjuvant chemotherapy response evaluation.⁷

Data extraction

Data extraction was independently carried out by 2 radiologists (O.B. and G.G.), each with 3 years of experience in breast imaging. Discrepancies were resolved through consensus, ensuring the reliability of the data. Extracted information included:

• Study characteristics: First author, country, publication year, study design.

• Population data: Number of patients, type of population (general, high risk, history of breast cancer), number of identified cancers.

• MRI details: Sequences in AB-MRI and FP-MRI protocols, scanning duration, number, and experience of MRI readers.

• Diagnostic performance metrics: Sensitivity, specificity, AUC, positive predictive value (PPV), and negative predictive value (NPV). For studies with multiple readers or various abbreviated protocol versions, these metrics were analysed separately for each.

Statistical analysis

We derived the 2×2 contingency table with the information in the included papers.

To account for the correlation between sensitivity and specificity, a bivariate model was used to calculate the summary estimates. Random effect models were used to calculate summary AUC and 95% CIs. Probability distributions for negative and PPVs are obtained by setting the prevalence range: 0.05-0.20. Subgroup analyses were conducted to investigate the difference between the AB-MRI and FP-MRI in terms of sensitivity, specificity, and AUC. Between-studies heterogeneity was assessed using the I² statistics, which quantifies the percentage of variation attributable to heterogeneity rather than chance. An I² below 50% was considered as an indicator of acceptable heterogeneity. In calculating the summary estimates, we considered that multiple estimates were from the same study.

All reported P-values were 2-sided, and P < .05 was considered statistically significant. Meta-analyses were carried out using the R software (R version 4.3.2).

Results

Literature research and study selection

The literature research through PubMed database yielded 367 studies and the application of our eligibility criteria led to the exclusion of 356 studies. Specifically, 336 studies were considered off-topic, and 20 studies were either reviews, meta-analysis or letters to the editor.

Ultimately, 11 studies met the inclusion criteria for this meta-analysis (Figure 1).

Figure 1.

Flowchart of the studies inclusion/exclusion criteria.

These comprised 10 retrospective studies, the majority of which (7/10) were retrospective cross-sectional studie,^8–14 2 retrospective cohort studies,^15,16 1 retrospective case-control study,¹⁷ and 1 prospective cohort study¹⁸ (Table 1).

Table 1.

The 11 selected studies.

Author (year)	Type of study	MRI field strength
Marquina Martínez (2019)	Retrospective cross-sectional study	1.5 T	AB-MRI sequences	AB-MRI acquisition time (min)	II AB-MRI sequences	II AB-MRI acquisition time (min)	III AB-MRI sequences	III AB-MRI acquisition time (min)
			Pre- and postcontrast axial T1W VIBRANT sequence acquired over 1 min, with and without subtraction MIP	N/S	MIP	N/S
			FP-MRI sequences	FP-MRI acquisition time (min)
			Axial T2-weighted (T2W) FSE SE-EPI diffusion-weighted, axial plane, with factors b = 0 and b = 600 Dynamic 3D FSPGR T1W study before and 1, 2, 3, 4, and 5 min after i.v. contrast MIP	N/S
Martin Drinković (2022)	Randomized cross-sectional comparative study	1.5T	AB-MRI sequences	AB-MRI acquisition time (min)	II AB-MRI sequences	II AB-MRI acquisition time (min)	III AB-MRI sequences	III AB-MRI acquisition time (min)
			One native and 4 postcontrast dynamic T1W volume imaging breast assessment (VIBRANT) sequences 3rd MIP	6-9
			FP-MRI sequences	FP-MRI acquisition time (min)
			Turbo spin-echo (TSE) T1WI Transversal TSE T2 images (T2WI) with fat suppression Sagittal TSE T2WI with fat suppression, Native and postcontrast dynamic T1W volume imaging breast assessment (VIBRANT) sequences 3rd MIP	25-35
Maryam Farghadani (2022)	Retrospective cross-sectional study	1.5 T	AB-MRI sequences	AB-MRI acquisition time (min)	II AB-MRI sequences	II AB-MRI acquisition time (min)	III AB-MRI sequences	III AB-MRI acquisition time (min)
			STIR precontrast T1WI fat sat (eg, THRIVE) 90 s postcontrast T1WI fat sat Subtraction images MIP	10
			FP-MRI sequences	FP-MRI acquisition time (min)
			T2WI STIR Pre contrast and 4 dynamic T1WI postcontrast sequences	40
Ko Woon Park (2020)	Retrospective case-control study	1.5 and 3T	AB-MRI sequences	AB-MRI acquisition time (min)	II AB-MRI sequences	II AB-MRI acquisition time (min)	III AB-MRI sequences	III AB-MRI acquisition time (min)
			T2WI TSE nonfat sat 3D T1WI fat sat precontrast and dynamic postcontrast (1 pre, 2 post contrast) Subtraction images and reversed subtraction MIP	9
			FP-MRI sequences	FP-MRI acquisition time (min)
			T1WI nonfat sat T2WI TSE fat sat DWI, ADC map 3D T1WI fat sat precontrast and dynamic postcontrast (1 pre, 6 post contrast) Subtraction images and reversed subtraction MIP Delayed T1 FS CE	23
Eun Sil Kim (2020)	Retrospective cross-sectional study	1.5 and 3T	AB-MRI sequences	AB-MRI acquisition time (min)	II AB-MRI sequences	II AB-MRI acquisition time (min)	III AB-MRI sequences	III AB-MRI acquisition time (min)
			Precontrast and first postcontrast T1WI MIP	N/S
			FP-MRI sequences	FP-MRI acquisition time (min)
			Fat-suppressed T2-weighted image (T2-WI) T1WI with 1 precontrast and 5 postcontrast series Subtraction images and reversed subtraction MIP	N/S
Soo-Yeon Kim (2022)	Retrospective cohort study	FP-MRI: 1.5T before 2016; 3T after 2016. AB-MRI: 3T	AB-MRI sequences	AB-MRI acquisition time (min)	II AB-MRI sequences	II AB-MRI acquisition time (min)	III AB-MRI sequences	III AB-MRI acquisition time (min)
			T2-WI One precontrast and 1 postcontrast T1WI Subtraction and MIP	8-9
			FP-MRI sequences	FP-MRI acquisition time (min)
			T2WI One precontrast and 5 postcontrast T1WI Subtraction and MIP	20-25
Marion E. Scoggins (2020)	Prospective cohort study	3T	AB-MRI sequences	AB-MRI acquisition time (min)	II AB-MRI sequences	II AB-MRI acquisition time (min)	III AB-MRI sequences	III AB-MRI acquisition time (min)
			Fat-suppressed T1W gradient-echo VIBRANT (1-2 min) First postcontrast fat-suppressed T1W gradient-echo VIBRANT (1-2 min) Subtraction images MIP images	N/S	Subtracted and nonsubtracted T1WI postcontrast (DCE)	N/S	T2WI TSE including fluid-bright and fat-only images DCE MIP	9,63
			FP-MRI sequences	FP-MRI acquisition time (min)
			T1W1 nonfat sat T2W2 fat sat T1WI 1 precontrast and 4 postcontrast fat sat Delayed postcontrast T1W1 DWI Subtraction images MIP	19,46
Isaac Daimiel Naranjo (2021)	Retrospective cohort Study	3 T	AB-MRI Sequences	AB-MRI Acquisition Time (min.)	II AB-MRI Sequences	II AB-MRI Acquisition Time (min.)	III AB-MRI Sequences
			Fat sat T1WI gradient-echo VIBRANT First postcontrast fat-sat T1WI gradient-echo VIBRANT Subtraction images MIP	10-15	Fat-sat T2-WI fast spin-echo Fat-sat T1WI gradient-echo VIBRANT First postcontrast fat-sat T1WI gradient-echo VIBRANT Subtraction images MIP	15-20
			FP-MRI Sequences	FP-MRI Acquisition Time (min.)
			Axial fat-sat T2WI fast spin-echo sequence axial non–fat-sat T1WI fast spin-echo 3D T1WI fat-sat gradient-echo sequence volume imaging for breast assessment (VIBRANT); before and after IV injection of with 3 postcontrast acquisitions at 60-s intervals Postprocessed subtracted images MIP	30-40
Zhenzhen Shao (2021)	Retrospective cross-sectional study	3T	AB-MRI sequences	AB-MRI acquisition time (min)	II AB-MRI sequences	II AB-MRI acquisition time (min)	III AB-MRI sequences
			Precontrast T1W First postcontrast T1W	N/S	Precontrast T1W First postcontrast T1W Axial DWI + ADC (with b values of 0 and 1000 s/mm2)	N/S
			FP-MRI sequences	FP-MRI acquisition time (min)
			Precontrast T1WI Fat-sat T2WI Axial DWI + ADC (with b values of 0 and 1000 s/mm2) Five postcontrast T1WI	N/S
Nasrin Ahmadinejad (2022)	Retrospective cross-sectional study	3T	AB-MRI sequences	AB-MRI acquisition time (min)	II AB-MRI sequences	II AB-MRI acquisition time (min)	III AB-MRI sequences
			FP-MRI sequences	FP-MRI acquisition time (min)
Peipei Chen (2022)	Retrospective cross-sectional study	3T	AB-MRI sequences	AB-MRI acquisition time (min)	II AB-MRI sequences	II AB-MRI acquisition time (min)	III AB-MRI sequences
			DWI T2WI	5,8	First postcontrast subtracted and MIP	3
			FP-MRI sequences	FP-MRI acquisition time (min)
			T2WI TSE fat sat TIWI TSE sequence DWI and ADC (b values including 0 and 800 s/mm2) T1WI dynamic gradient-echo fat sequence	17

Abbreviatios: AB-MRI = abbreviated magnetic resonance imaging; ADC = apparent diffusion coefficient; DWI = diffusion-weighted imaging; FP-MRI = full diagnostic protocol MRI; MIP = maximum intensity projection; T1W = T1-weighted; T2W = T2-weighted; VIBRANT, volume image breast assessment; THRIVE, T1-weighted high-resolution isotropic volume examination; TSE, turbo spin-echo.

Patient demographics and study characteristics

Retrospective studies included 3321 patients in the screening or diagnostic setting and 1506 patients with a positive history of breast cancer; the prospective study included a total of 73 patients.

Technical aspects and MRI protocols

Technical details and the sequences included in both the abbreviated and standard protocol of selected studies are reported in Table 1.

Five out of 11 studies compared one abbreviated protocol composed of at least one unenhanced and one contrast-enhanced sequence to the standard full diagnostic protocol.

The remaining 6 studies proposed either 2 or 3 different abbreviated protocols, composed again at least of one unenhanced and 1 contrast-enhanced sequence with some studies also integrating Diffusion-weighted imaging (DWI) with apparent diffusion coefficient (ADC) mapping and/or maximum intensity projection (MIP).

The mean acquisition time recorded for AB-MRI was 8.37 min, in contrast to 28.75 min for FP-

Statistical analysis

The statistical analysis showed a significant difference in terms of sensitivity between the 2 protocols (95.0% FP-MR vs 86.0% AB-MRI, P = .005), but not for the specificity (P = .50) (Table 2).

Table 2.

Summary for sensitivity and specificity of the studies, with the N^∧ being the estimated number of readers, considering that there were more than one in each study.

	AB-MR		FP-MRI
	N ∧	Summary (95% CI), $I^2$	N ∧	Summary (95% CI), $I^2$	P-value
Sens	14	86% (75%-93%), 34.2%	14	95% (84.0%-98%), 37.5%	.005
Spec	14	86% (67%-95%), 96.9%	14	86% (68%-94%), 97.2%	.50

Forest plots for sensitivity and specificity are shown in Figure 2.

Figure 2.

Forest plots for sensitivity and specificity. FP-MRI appears to have a significantly higher summary sensitivity compared to AB-MRI, while the summary specificity is similar between the 2 methods. Abbreviations: AB-MRI = abbreviated breast MRI; FP-MRI = and full protocol MRI.

Furthermore, FP-MRI achieved a significantly higher summary AUC than the AB-MRI (92.0%, 95% CI 88.6%-95.2% and 85.9%, 95% CI 81.9%-90.2%, respectively, P = .0007).

We obtain a distribution of NPV defined by a mean of 97.9% and 96.2% for FP-MRI and AB-MRI, respectively. A distribution of PPV is defined by a mean of 32.3% and 32.8% for FP-MRI and AB-MRI, respectively (Figure 3).

Figure 3.

Projected predictive value in AB-MR and FP-MRI by setting the prevalence range: 0.05-0.20. Abbreviations: AB-MRI = abbreviated breast MRI; FP-MRI = and full protocol MRI.

Discussion

In this meta-analysis, 11 studies were aggregated to compare the diagnostic accuracy of AB-MRI and FP-MRI in both screening and diagnostic settings for breast cancer.

The results show a statistically significant higher sensitivity in FP-MRI (92.4%) compared to AB-MRI (82.3%), with a P-value of .005. This indicates that FP-MRI is more effective in identifying breast cancer cases, which is crucial for early detection and treatment. The higher sensitivity of FP-MRI can be attributed to its comprehensive nature, capturing a broader range of diagnostic information.

Conversely, no significant difference was found in specificity between the 2 protocols (P = .50), suggesting that both modalities are equally effective in correctly identifying patients without breast cancer.

The higher AUC for FP-MRI (P = .0007) compared to AB-MRI underlines the overall diagnostic superiority of the full protocol. The AUC, a measure of a test’s ability to discriminate between conditions, indicates that FP-MRI more effectively distinguishes between cancerous and non-cancerous cases.

Both protocols demonstrated high NPV, with FP-MRI slightly higher (97.9%) than AB-MRI (96.2%). This high NPV is critical in clinical settings as it indicates the likelihood of a patient not having breast cancer when the test is negative, thus reducing the risk of missed diagnoses.

The PPV was 32.3% for FP-MRI and 32.8% for AB-MRI. Similar PPVs suggest that, when a test is positive, the likelihood of actual breast cancer presence is moderately low for both protocols. This could be influenced by the prevalence of breast cancer in the study populations and underscores the need for cautious interpretation of positive results.

In this meta-analysis, the primary endpoint was the aggregate accuracy of the abbreviated MRI protocol; hence, studies conducted in the diagnostic and screening settings were not separately analysed. Of the included studies, those conducted in the screening setting constituted a minor proportion (5/11), and all of them included high-risk patients with a lifetime risk of developing breast cancer of 20%-25% or higher. This cohort included individuals with BRCA1 and BRCA2 mutations, as well as those with familial or personal histories of breast or ovarian cancer, and those with a history of thoracic radiotherapy. Consequently, the study population was not representative of the general population, but rather, it was a distinct cohort with an intrinsically elevated risk for breast cancer development.

Interestingly, the lowest sensitivity for the AB-MRI protocol was reported by Kim et al.¹¹ The study involved a multireader panel of 5 fellowship-trained radiologists whose breast MRI interpretation experience ranged from 3 to 11 years, with an average of 7.4 years. This mean experience is somewhat lower than the median of 9.8 years reported in the papers selected in the present meta-analysis, yet not significant to explain such difference also considering that the radiologists in their study had to pass an AB-MR Reader Training and Certification Test before participation. Their AB-MRI protocol was in line with the ones of the other studies, consisting of a precontrast T1-weighted image (T1WI), first postcontrast T1WI, and an MIP.

Most of the studies, in fact, proposed an AB-MRI protocol composed of at least a precontrast and a postcontrast T1WI, with some timing variation, for example, Farghadani et al¹³ utilized 90 s postcontrast T1WI fat sat whereas Drinković et al¹² included 4 postcontrast dynamic T1WI, and an MIP. As the timing of the postcontrast sequences has an important influence on the overall duration of the whole examination, most authors tried to reduce the number of these sequences, including only the ones valued as key sequences to achieve an optimal diagnosis.

More variable was the employment of T2WI used by 5 authors,^10,15–18 and of DWI proposed only by 2 authors.^8,10

Particularly, Kim et al assessed that simulated AB-MRI with single first postcontrast images employed in the study, showed a lower sensitivity than FP-MRI, attributing this difference to the absence of kinetic information and to the fact that the more suspicious irregular margins and internal or heterogeneous enhancement were observed in the delayed phase of the standard protocol, accounting for a higher number of FN results in the AB-MRI.

The rationale behind the use of only the first postcontrast T1WI is that tumours, having an extensive neo-angiogenesis, typically present early contrast enhancement and, different studies promoted an AB-MRI protocol including only the first postcontrast T1WI.

For instance, Mango et al¹⁹ demonstrated that the first postcontrast AB-MRI had high sensitivity for detection of breast cancers.

Furthermore, slowly enhancing malignancies such as low-grade ductal carcinoma in situ or invasive lobular carcinomas might be overlooked by AB-MRI, which lacks delayed postcontrast phases, leading to missed diagnoses. Conversely, late-enhancing benign lesions could result in a higher rate of false-positives, particularly for BI-RADS 3 category, that is, Kim et al,¹⁵ showed a twice as high rate of lesions categorized as BI-RADS 3 in FP-MRI (12%) compared to AB-MRI (5%).

To address these challenges while preserving AB-MRI's advantages, incorporating alternative postcontrast images may be beneficial. Shao et al⁸ improved the diagnostic accuracy of their AB-MRI consisting of a precontrast and a single first postcontrast T1WI though the addition of DWI (b: 0/1000 s/mm²) and ADC map: the sensitivity increased from 95.6% without an ADC to 98.0% with an ADC, the PPV from 88.9% to 89.1%, and the NPV from 93.4% to 96.9%.

The incorporation of T2WI into AB-MRI protocols is still under discussion. While T2WI may not directly impact cancer detection, Naranjo et al¹⁶ suggested that it enhances lesion conspicuity and specificity, and its addition did not compromise the sensitivity or accuracy of AB-MRI compared to standard protocols.

Regarding the possibility of using an unenhanced protocol Chen et al¹⁰ uniquely investigated an MRI protocol without contrast agents using DWI and T2WI among the studies reviewed. This unenhanced protocol, compared alongside an AB-MRI with postcontrast sequences, showed a considerable reduction in scan and interpretation times relative to the FP-MRI. However, the unenhanced protocol exhibited lower sensitivity, particularly for lesions ≤10 mm. This observation aligns with findings by Pinker et al²⁰ and Rotili et al,²¹ noting diminished sensitivity for smaller tumours (namely, for lesions ≤10 mm).

The detection of small lesion remains a key issue as the majority of lesions missed across the reviewed studies were small in size, regardless of the AB-MRI sequence used. Kim et al,¹⁵ reported 5 false-negative results on AB-MRI with average size of 106 cm which were detected on FP-MRI due to delayed imaging features. Shao et al⁸ reported non–mass enhancement as the most commonly overlooked cancer type, with all missed mass-enhancement cancers being under 2 cm. Additionally, the inclusion of an ADC value in the AB-MRI protocol reduced the number of missed lesions.

The abbreviated protocol’s reduced acquisition and interpretation times, which enhance accessibility and patient experience, were confirmed by the studies included in our analysis, except for Shao et al.’s study,⁸ which did not assess reading times. The inclusion of ADC values, while increasing diagnostic accuracy, also increased interpretation time. Additionally, the experience level was a significant factor in interpreting DWI images, as demonstrated by Rotili et al,²¹ where less experienced reader showed lower sensitivity. Based on the results of the aforementioned studies, an abbreviated protocol that includes ultrafast dynamic contrast-enhanced MRI, T2-STIR–weighted imaging, and DWI with the corresponding ADC map—for a total duration of 12 min—is under evaluation: such protocol aims to balance diagnostic accuracy and efficiency, making AB-MRI a more feasible option for widespread screening applications.

The ultrafast protocol may allow rapid acquisition of contrast-enhanced images, crucial for detecting enhancing lesions quickly and efficiently, with the T2-STIR weighted imaging providing high-contrast resolution of breast tissue (which is particularly useful in identifying lesions in dense breast tissue) and the DWI and ADC map that add another layer of diagnostic capability by assessing tissue cellularity and distinguishing between benign and malignant lesions.

Our study acknowledges limitations, including a small dataset and heterogeneity in the populations studied, study types, and reported protocols. The considerable heterogeneity observed in the populations studied, the types of studies selected, and the MRI protocols reported could introduce variability and affect the consistency of the results. Consequently, further research is necessary to determine the optimal sequences for an AB-MRI that considers the benefits and limitations of various sequences.

Conclusion

This meta-analysis of 11 studies showed that AB-MRI had an overall acceptable diagnostic performance in the detection of breast cancer.

While a significant difference in sensitivity between AB-MRI and FP-MRI was observed, specificity remained comparable.

To validate AB-MRI’s diagnostic accuracy and ascertain its cost-effectiveness, acceptability, and practicality in both diagnostic and screening contexts, large-scale, multicenter prospective trials are necessary. Future research should aim to refine AB-MRI protocols, clarifying its indications and identifying the patient populations that would most benefit. This includes examining the potential role of machine learning in improving lesion detection and characterization within AB-MRI frameworks.

Our findings indicate that while FP-MRI remains the gold standard for breast cancer detection due to its higher sensitivity and comprehensive diagnostic capability, AB-MRI emerges as a viable, quicker, and potentially more accessible option, especially in scenarios where the extensive duration and costs of FP-MRI are limiting factors. The shorter acquisition time and comparable specificity of AB-MRI render it a promising approach for screening, particularly in high-risk groups. However, its relatively lower sensitivity compared to FP-MRI indicates limitations as a standalone diagnostic tool.

It is crucial to note that an effective screening tool for breast cancer must possess high sensitivity to ensure the identification of all potential cases for further evaluation. Diagnostic tests, on the other hand, require high specificity to minimize unnecessary biopsies. Given this, the introduction of AB-MRI as a screening tool is particularly supported for patient categories where there is no consensus for screening with FP-MRI, yet mammography, even when enhanced with digital breast tomosynthesis, is not sufficiently sensitive, especially in women with dense breasts.^10,22

The shorter acquisition time and comparable specificity of AB-MRI may render it a promising approach for screening women with an intermediate lifetime risk of developing breast cancer: this includes women with dense breasts, a previous history of breast cancer, atypical ductal hyperplasia, or other lesions with uncertain malignant potential. For these specific patient groups, there are no formal recommendations for periodic breast MRI screening. AB-MRI, with its significantly higher sensitivity, presents a viable compromise, addressing some of the limitations associated with FP-MRI screening.

Future research should focus on optimizing AB-MRI protocols to enhance sensitivity while maintaining the benefits of shorter acquisition times, and on standardizing these protocols to improve consistency and reliability across different clinical settings.

Furthermore, a comprehensive understanding of patient-specific factors—such as mammographic density, genetic predispositions, and individual risk profiles—could refine the selection process between AB-MRI and FP-MRI, thereby enhancing breast cancer screening and diagnostic practices by a multidisciplinary approach, integrating radiological, pathological, clinical, and genetic data.²³

Acknowledgements

This work was partially supported by the Italian Ministry of Health with Ricerca Corrente and 5 × 1000.

Conflicts of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.