Evaluation of Diagnostic Performance of Automatic Breast Volume Scanner Compared to Handheld Ultrasound on Different Breast Lesions: A Systematic Review

Objective: To compare the diagnostic performance of the automatic breast volume scanner (ABVS) against the handheld ultrasound (HHUS) in the differential diagnosis of benign and malignant breast lesions. Methods: A systematic search and review of studies involving ABVS and HHUS for breast cancer screening were performed. The search involved the data taken from Scopus, PubMed, and science direct databases and was conducted between the year 2011 to 2020. The prospective method was used in determining the inclusion and exclusion criteria while the evidence level was determined using the BI-RADS categories for diagnostic studies. In addition, the parameters of specificity, mean age, sensitivity, tumor number, and diagnostic accuracy of the ABVS and HHUS were summarized. Results: No systematic review or randomized controlled trial were identified in the systematic search while one cross-sectional study, eight retrospective studies, and 10 prospective studies were found. Sufficient follow-up of the subjects with benign and malignant findings were made only in 10 studies, in which only two had used ABVS and HHUS after performing mammographic screening and MRI. Analysis was made of 21 studies, which included 5448 lesions (4074 benign and 1374 malignant) taken from 6009 patients. The range of sensitivity was (0.72–1.0) for ABVS and (0.62–1.0) for HHUS; the specificity range was (0.52–0.98)% for ABVS and (0.49–0.99)% for HHUS. The accuracy range among the 11 studies was (80–99)% and (59–98)% for the HHUS and ABVS, respectively. The identified tumors had a mean size of 2.1 cm, and the detected cancers had a mean percentage of 94% (81–100)% in comparison to the non-cancer in all studies. Conclusions: The evidence available in the literature points to the fact that the diagnostic performance of both ABVS and HHUS are similar with reference to the differentiation of malignant and benign breast lesions.


Introduction
Breast disease is common among modern women. It is also one of the leading diseases that threatens the physical health of women. The American Cancer Society predicted that the United States would top the ranking, with 276,480 women with breast cancer in 2020 in the country alone, which accounted for 30% of all cancer patients. With the death of 42,170 of these women, it ranked in second place by accounting for 15% of cancer deaths [1]. In Malaysia, breast cancer is the leading type of cancer, which accounted for 34.1% of all cancer cases in the female population. The diagnosis of a total of 21,634 female breast

Selection Criteria
The criteria for selection depended on the PRISMA Statement [21]. The search mainly focused on the mapping of existing literature on automated breast ultrasound in contrast to handheld ultrasound in medicine, biochemistry, genetics, molecular biology, and the health professions. The search then narrowed down to the medical field. The search span was made between 2011-2020. All articles before 2011 were excluded. The search was not limited to any specific countries, therefore there was no exclusion in this option. The following are the inclusion criteria: (1) both ABVS and HHUS were used in breast lesion diagnosis; (2) the ABVS method was financially accessible; (3) the study population was made up of a minimum of 20 patients; (4) follow-up of histologic analysis (surgery or biopsy), and clinical/imaging for a minimum of 1 year, and unchanged lesions were viewed as pathologically benign. Screening of the relevant literature's abstracts and titles were conducted, and inspection of the full texts was performed by two researchers independently in determining the inclusion of selected articles in the analysis. A consensus was used in resolving any conflict between the two researchers. At this stage, a total of 311 research articles were excluded while 414 records were extricated.

Data Extraction
Data collection was made regarding the year of publication, the country in which the examination was performed, the objectives, study design, number of participants, screening methods of assessment, patients' mean age, and the number of lesions.

Quality Assessment
Conference papers, original research articles, and review papers became the basis of this study. A thorough check was made on all duplications in maintaining the nature of the review. In ensuring the relevance and quality of the academic literature included in the review process, detailed examination of the abstracts of articles was conducted in the process of analysis and purification. Next, careful assessment of each research paper was performed. To limit the research only to English-published papers, the subsequent exclusion criterion was therefore used. Therefore, three articles in languages other than English have not been included in the study. In addition to that, the filtration of duplicate records resulted in the removal of 65 more articles. The assessment of each article based on the inclusion and exclusion criteria above resulted in the selection of 21 articles. The exclusion and inclusion of the literature at every stage (PRISMA Statement) is shown in Figure 1. Assessment of the methodological quality was made by two independent reviewers, and to resolve any dispute between the reviewers, mutual suggestion was used. For the inclusion of studies regarding 'diagnostic accuracy', the QUADAS-2 (Quality Assessment of Diagnostic Accuracy Studies-2) tool was used which involved four domains including 'index test', 'reference standard', 'patient selection', and 'flow and timing' [22] ( Table 1). Evaluation was made of each domain with regard to its risk of bias (low, high, or unclear), and the initial three domains were identified with regard to their applicability. In general, a study that is viewed as "low" in all domains regarding its applicability or bias is deemed appropriate for an overall judgment of "low concern regarding applicability" or "low risk of bias" for the study. However, judgement of a study is made as having "concerns regarding applicability" or "at risk of bias" if viewed as "unclear" or "high" in one or more domains. Dispute between the two reviewers in assessing the quality of the study was settled through discussion ( Figure 2).

Search Strategy and Study Selection
Based on the search strategy, 726 records were discovered from the electronic databases when, due to duplication issues, 311 items were then discarded. The screening of titles and abstracts based on the inclusion criteria resulted in the exclusion of another 320 items.  Figure 2). utilized among three studies, EUB-8500 scanner (Hitachi Medical, Tokyo, Japan) used in two study while Aplio 80 (Toshiba, Tokyo, Japan) just in one study.

Characteristics of the Study and Quality Assessment
Among 14 studies, the sensitivity range was (0.72-1.0) for ABVS and (0.62-1.0) for HHUS, while the specificity range was (0.52-0.98) for ABVS and (0.49-0.99) for HHUS. Among the 11 studies, the accuracy range was (59-98)% and (80-99)% for ABVS and HHUS, respectively. ABVS and HHUS were utilized in 21 studies, thus giving unbiased, extractable data in terms of diagnostic accuracy. For the 21 studies on 'diagnostic accuracy', the results of histopathology assessment were the standard of reference. 'Lesion by lesion' is the rater approach chosen in all studies. The raters evaluated each lesion detected for malignancy using the BI-RADS. Table 1 provides the summary of the details of the study characteristics. As demonstrated by the QUADAS-2 tool, most of the studies (15 out of 21) had a rather high methodological quality (Table 2). However, due to the use of case-control configuration in this study, one study was judged as high risk [23]. As a result of imprecise review of HHUS and ABVS images, the same radiologist judged five studies as unclear risk, with the use of the blinding method.

Age Distribution
The report of 18 studies showed an age range between 11 to 82 years old, with the overall age range exceeding 30 years old in every study. Information on the median age was provided in three studies (49, 49, and 52 years), with age range between 32 to 82 years [24][25][26] (Table 2).

Breast Lesion Diagnosis Using the Categories of BI-RADS
Two studies separated the HHUS and ABVS outcome of breast screening according to the BI-RADS 4 and BI-RADS 5 classification [25,27]. Other studies assessed the results of breast tissue according to BI-RADS 1-BI-RADS 5. Females with BI-RADS 4 and BI-RADS 5 breast categories demonstrated the largest proportion of breast cancer diagnosed through ultrasound screening. Depretto et al. analyzed four carcinomas distinguished by (BI-RADS 4) breast tissue, and 130 cases with breasts in categories (BI-RADS 1 and BI-RADS 2) [28], and 29 malignancies were diagnosed in type 4 and type 5 BI-RADS breast tissue in Jia et al., while two carcinomas were found in three breasts [23].
Using the ABVS, the discovery of 51 circumscribed solid nodules (BI-RADS 3) were made in 42 women. The HHUS exhibited five of these as complicated cysts. The primary HHUS missed five nodules. However, the detection of five BI-RADS 3 solid nodules, one BI-RADS 2 implant rupture, and one BI-RADS 4 distortion was made by HHUS, which were all missed by ABVS. Altogether, 78 lesions were found in 340 women, 71 (91%) of the detected nodules were discovered by ABVS while 68 (87.2%) of the detection was made by the primary HHUS [24]. Niu et al. included 599 masses detected in 398 women (which included solid and cystic masses). The classification of 359 masses by the HHUS and ABUS as category 2 or 3 indicated these as benign masses. The MRI classified two of these masses as category 4 or 5 [29].
It was found at the completion of the study that 496 (83%) of the 599 masses were benign while the remaining 103 (17%) were malignant, with the one-year follow-up information or pathological outcomes as the reference standards. As indicated by the BI-RADS classification for every unit, 258 units (258/320, 80.63%) had BI-RADS classification 1-  In comparison to HHUS, ABVS had irregular nodules of (p < 0.001), distortions of (p < 0.034), and over diagnosed shadowing of (p < 0.01). 59.7% of the women mentioned that if they had a choice, they would have chosen HHUS. 10.6% of the women experienced severe pain from the use of ABVS. The use of ABVS in lesion detection is as good as HHUS. ABVS had a higher recall rate and lower positive predictive value. This could end up in greater follow-ups, and greater anxiety for the women. Had they been given the choice, more than 50% of the women would have preferred HHUS.

Depretto et al., 2020 Italy
Study design = Retrospective study Objective There was a significant difference in assigning the BI-RADS categories (p < 0.05), although there was no difference in dichotomic assignment between 2 readers (p = 0.5).
A substantial agreement was achieved between ABUS and HHUS in monitoring of women with breast cancer history.
In particular, the ABUS could be used in first-level monitoring of intermediate risk women since it could recognize all cancers detected by HHUS.

Jia et al., 2020 China
Study design = Cross-sectional Objective

•
To determine, both in combination with mammography and separately, the diagnostic performance of the automated breast ultrasound system (ABUS) and the handheld ultrasound (HHUS) in dense breasts Chinese women. The combination of the HHUS with mammography produced a 0.92 area under the curve, while a combination of ABUS with mammography produced one at 0.93. An agreement that is statistically significant in breast cancer detection between HHUS and ABUS was observed (percent agreement = 0.94, κ = 0.85). As adjuncts to mammography, both the HHUS and ABUS can substantially increase the rate of breast cancer detection in dense breasts women; a strong correlation exists between them. With various benefits of the ABUS over HHUS, for instance reproducibility and less operator dependence, and the commonness of dense breasts, the use of ABUS in the early detection of breast cancer, especially in areas that have limited resources, has shown great potential.       diagnostic performance might not be significantly improved through coronal reconstruction [30]. In the study of Zhang et al., a suggestion was made that ABVS independent value limitation in the differential diagnosis is caused by the low sensitivity (37.0%) of the retraction phenomenon on the coronal plane [32]. The limitations in our study include the substantial dominance of Asian reports, with 15 out of 21.
Variations might have occurred due to the uneven geographic distribution since there are breast cancer differences regarding region and ethnicity between non-Asian and Asian women. Besides, no indication was made in any of the studies included regarding an image quality control statement, which should thus be noted as a variable that is unaccounted for, both for ABVS and HHUS. Third, based on our references, no investigation has made use of Doppler and elastography ultrasound in HHUS, in contrast to the practice in clinical reality. Finally, publication bias might have been prompted since evaluation was made only of articles written in English. Therefore, our reviewed articles may have underestimated the diagnostic performance of HHUS.

Conclusions
In relation to malignant and benign breast lesion differentiation, ABVS diagnostic performance based on the evidence available in the literature is similar to that of HHUS. However, ABVS can offer new diagnostic information. ABVS may help to distinguish between real lesions. This technique is feasible for clinical applications and is a promising modality in breast imaging. Nevertheless, since this review of articles was conducted on various studies, most of which were obtained from a single geographical region, further studies are hence required before the generalization of this conclusion can be made. More sound research associating the diagnostic performance of ABVS and mammography/MRI is anticipated and required.