Pathological Results (Benign/Malignant) Can be Inferred from Macroscopic Findings for Breast Vacuum Associated Biopsy Specimens

Background: If the histopathological results could be inferred from the macroscopic ndings for breast biopsy specimens, false-negative needle biopsies due to poor specimens could be prevented. However, no research on this has been reported to date. Methods: Biopsy specimens from 101 patients who underwent mammary gland Vacuum Associated Biopsy (VAB) were photographed with a smartphone, and the relationships between the macroscopic ndings (presence/absence of turbidity, surface properties, presence/absence of white spots, and characteristic ndings) and the pathological results (benign/malignant) were examined. Results: A signicant difference was observed with regard to the presence/absence of turbidity: malignancy was detected in 33/37 (89%) specimens with turbidity and in 2/47 (4%) cases without turbidity (p<0.001). A signicant difference was also observed regarding the surface properties: malignancy was detected in 14/70 (19%) smooth specimens and in 24/29 (83%) rough specimens (p<0.001). Also, malignancy was detected in 11/13 (85%) specimens with white spots, and the difference was signicant (p<0.001). Conclusions: The pathological results (benign/malignant) were able to be predicted from the macroscopic ndings for the breast VAB specimens. These ndings warrant performance of reproducibility and prospective studies at multiple facilities.


Introduction
Needle biopsy is an essential modality for diagnosis of breast disease [1 ~ 5]. In recent years, many small lesions have been detected due to mammography-combined screening for breast cancer and improvements in breast ultrasound equipment, and the target lesions of needle biopsy have also become smaller. The success or failure of needle biopsy is often in uenced by the skill of the operator, but one of the causes of false-negatives is that the lesion was not properly biopsied [6 ~ 8]. In particular, in the case of small lesions, it is more likely that biopsy will not be performed properly. Calci ed lesions can be con rmed by radiography of biopsy specimens, but there is no con rmation method for lesions without calci cation [8 ~ 11]. On the other hand, if the macroscopic ndings of a specimen obtained by needle biopsy match the predicted histopathological ndings, it can be judged that the lesion was properly biopsied. However, there are no reports on the macroscopic ndings for needle biopsy specimens. This retrospective study investigated whether the histopathological results can be inferred from the macroscopic ndings for breast Vacuum Associated Biopsy (VAB) specimens. It was approved by the Ethics Committee of Tokushima University School of Medicine. For this type of study formal consent is not required.

Subjects And Methods
Macroscopic study of normal breast VAB specimens Five mammary glands totally resected at Higashi Tokushima Medical Center were immediately subjected to VAB of a site that was thought to be normal. The obtained VAB specimens were photographed and macroscopically con rmed to be normal. The photographic procedure was as follows. A smartphone (iPhone 7) was equipped with a 3x macro-lens. The specimen was illuminated with a commercially available illumination device (MAG-LITE SOLITAIRE LED FLASHLIGHT, 47 LUMENS, MAG INSTRUMENT,   INC.; USA) at an angle of about 30 degrees to enable observation with tangential light (Fig. 1). The ndings were judged by two surgeons and nalized by discussion.
As a result, it was judged that the normal mammary gland tissue in the VAB specimen was transparent and had a smooth and shiny surface. Therefore, our current study of macroscopic ndings focused on (1) whether there was loss of transparency (that is, whether there was turbidity), (2) the surface properties, (3) the presence/absence of white spots, and (4) the presence of characteristic ndings. The signi cant ndings were selected, but when it was clear that there were mixed ndings, the specimen was judged as mixed.

Subjects
From January 2018 through December 2018, VAB was performed for 102 patients at Tokushima Breast Care Clinic. VAB was performed under ultrasound-guided local anesthesia by a single surgeon using a breast biopsy device, Celero-12 (Hologic Ltd.; Manchester, UK). Based on the image ndings and histopathological results, all but one case was judged to have been properly biopsied. The reasons for performing VAB were as follows: in 25 cases, ne needle aspiration biopsy (FNA) was suspicious for malignancy; in 44 cases the FNA was indeterminate; in 17 cases, imaging was suspicious for malignancy, but FNA indicated benign; in 5 cases, imaging was suspicious for malignancy, but FNA was not performed; in 5 cases, MRI examination performed before breast cancer surgery indicated a new lesion, and 2nd -look US detected lesions; and in 5 breast cancer cases, VAB was required for subtype determination. The age range was 32 to 81 years (mean, 51.7 years), and the lesion size was 3.4 to 27 mm (mean, 11.4 mm).

Macroscopic ndings for VAB specimens
The VAB specimens were immediately photographed with a smartphone. The imaging method was the same as that described above for surgical specimens. The macroscopic ndings were later evaluated by two doctors (one surgeon and one radiologist), and a consensus evaluation was reached. The evaluated features were: (1) presence/absence of turbidity (turbidity, no turbidity, mixed, di cult to evaluate); (2) surface properties (smooth (smooth and/or shiny), rough, mixed, di cult to evaluate); (3) presence/absence of white spots; and (4) characteristic ndings (papillary structure, circular granular structure, contraction, edema, red cord-like structure, presence of mucus).
Examination for relationship between VAB specimen macroscopic ndings and pathological results
For the benign specimens, 73% of intraductal papillomas (IDP) and 67% of ductal hyperplasia (DH) showed no turbidity, and none of the broadenomas (FA) showed turbidity. The differences were not signi cant (Table 2).
Relationship between the macroscopic ndings and pathological results (benign/malignant) with consideration of turbidity and the surface properties (rough/smooth) Evaluation was performed for 100 specimens after excluding one specimen that consisted of a mucus mass, 21/22 (96%) turbid specimens with a rough surface were malignant, whereas malignancy was found in only 4% of specimens that were not turbid and had a smooth surface (Table 3). Sixty-one percent (22/36) of the turbid specimens had a rough surface, while 96% (45/47) of the specimens without turbidity were smooth, indicating a clear relationship between the presence/absence of turbidity and the surface properties (data not shown).
Relationship between white spots and pathological results (benign/malignant) All 101 specimens were evaluated for white spots. White spots were observed in 13 specimens (Fig. 3a), of which 11 (85%) were malignant. The remaining 88 specimens had no white spots, and 29 (33%) were (benign/malignant) The pathological results for VAB specimens were divided into two categories, malignant and benign, and examined for a relationship with the macroscopic ndings. In addition, benign disease specimens were examined for a relationship between turbidity/no turbidity and the histopathological result. Furthermore, the characteristic ndings were examined for a relationship with the histopathological result. Statistical testing used the χ2 test, and p < 0.05 was de ned as signi cant.
Relationships of the characteristic ndings with pathological results (benign/malignant) and the histopathological results Table 4 shows the characteristic ndings and the histopathological results. As the main characteristic ndings, 37 specimens had a papillary structure (Fig. 3b), 33 specimens had a circular granular structure (Fig. 3c), 12 specimens showed contraction (Fig. 3d) and 9 specimens showed edema (Fig. 3e). With regard to the relationship between the characteristic ndings and pathological results (benign/malignant), malignancy was found in 42% (14/33) and 50% (6/12), respectively, of the circular granular structure specimens and the contracted specimens. With regard to the histopathological result, more than half of the IDP specimens had a papillary structure, and 32% were malignant. FA specimens were characterized by edema and red cord-like structures (Fig. 3e). Mucus was present in only one specimen, and that case was mucinous cancer.

Discussion
Today, needle biopsy is an essential technique for preoperative diagnosis of breast cancer. One of the major causes of false negatives in needle biopsy diagnosis is that the lesion was not properly biopsied [6 8]. If the macroscopic ndings of the needle biopsy specimen match the expected histopathological results, it can be judged that the needle biopsy was properly performed. Our literature search found that Rosen p.p. performed detailed examination of the gross ndings for excised fresh specimens [12], but we found no reports of studies of needle biopsy specimens [9]. Our present report is thus the rst to examine needle biopsy specimens in detail.
This was a retrospective study. Specimens were photographed with an iPhone. This is because the number of pixels of the iPhone camera, 12 million, was judged to be su cient for documenting the specimens properties, while the camera can be easily used in ordinary facilities. A commercial, readilyavailable illumination device was used for tangential lighting. Rosen p.p. had used tangential light to observe excised specimens, and we judged that tangential light was suitable for accurately observing the surface properties of our needle biopsy specimens. In addition, we examined VAB specimens because larger specimens can be evaluated compared to core needle biopsy (CNB) specimens.
As macroscopic ndings, we examined only the 1) presence/absence of turbidity, 2) surface properties, 3) presence/absence of calci cation and 4) characteristic ndings. Turbidity was examined because it has been thought to be associated with an increase in cell components. The surface properties were examined because they have been thought to re ect the in uence of tumors on the stroma. White spots were examined because they are thought to represent microcalci cations. Our terminology for the characteristic ndings was based on the terminology used for gastrointestinal endoscopic ndings [13].
The ndings were retrospectively reviewed and classi ed. We did not evaluate the color of our VAB specimens because Rosen p.p. reported that, for their excised specimens, even invasive ductal carcinoma (IDC) exhibited a range of colors (white, gray and yellow) [12]. Regarding the pathological results, because the number of evaluated cases was small, all the cases of breast cancer were considered malignant, and we did not examine for a relationship with the histopathological result. On the other hand, for cases of benign disease, we did examine the presence/absence of turbidity and the characteristic ndings for relationships with the histopathological result.
Our results indicate that malignancy was signi cantly more frequent in specimens with turbidity and a rough surface. Furthermore, the malignancy rate was 96% for specimens with both turbidity and a rough surface. In addition, only 13% of our specimens had white spots due to microcalci cation, but their rate of malignancy was signi cantly high. Collectively, specimens showing turbidity, specimens with a rough surface, and specimens with white spots were signi cantly more likely to be malignant. Also, the presence/absence of microcalci cations was able to be con rmed by visual observation.
Next, specimens with a circular granular structure as a characteristic nding had a malignancy rate of 42%, while the rate was 50% for specimens with contraction. Although papillary structures are common in IDP, 32% were malignant. In addition, edema and red cord-like structures were found to be characteristic of FA. In Rosen p.p. study using fresh samples, radial sclerosing lesions, which are benign disease, had a retracted center with white streaks, while cases with adenosis had abundant calci cations and appeared to be gritty. On the other hand, for IDC they wrote that "the appearance of the cut surface very considerably depended on the composition of the tumor," etc. [12]. Excised specimens and VAB specimens differ not only in size: it is predicted that excised specimens often include normal mammary gland tissue, and the macroscopic ndings are also expected to differ between these two types of specimen. In the future it will be necessary to investigate VAB specimens for a relationship between their macroscopic ndings and histopathological results.
This study has a number of limitations. (1) The results of the dedicated imager and smartphone should be compared and examined, but we did not do this. (2) Turbidity and surface properties are subjective evaluations, and a multi-institutional reproducibility study is needed. (3) It is necessary to investigate whether CNB specimens can also be used. (4) This was a retrospective study, and a prospective study needs to be carried out.
Although the number of evaluated VAB specimens was small, it was found that the pathological results (benign/malignant) can be predicted from the macroscopic ndings. In the future, it will be necessary to conduct reproducibility and prospective studies at multiple institutions.

Conclusions
This study indicate that specimens showing turbidity, specimens with a rough surface, and specimens with white spots were signi cantly more likely to be malignant. The pathological results (benign/malignant) were able to be predicted from the macroscopic ndings for the breast VAB specimens.  Tables   Due to technical limitations, table 1 to 4 is only available as a download in the Supplemental Files section. Figure 1 Macroscopic ndings for a VAB specimen of a normal mammary gland. The normal mammary gland is rather transparent, and the surface appears smooth under tangential light. (normal mammary gland). Figure 2 a: Macroscopic ndings for a VAB specimen; turbidity. Decreased light transmission indicates the presence of turbidity. In this case, bleeding is also present and a papillary. Bleeding is also present. A papillary structure is seen (non-invasive ductal carcinoma). b: Macroscopic ndings for a VAB specimen; no turbidity. The transparency is the same as that of a normal mammary gland (intraductal papilloma). c:

Figures
Macroscopic ndings for a VAB specimen; rough surface. Fine-sand-like, light-re ecting spots are observed under tangential light, and surface irregularities are present overall (invasive ductal carcinoma). Figure 3