The differentiation of the character of solid lesions in the breast in the compression sonoelastography. Part I: The diagnostic value of the ultrasound B-mode imaging in the differentiation diagnostics of solid, focal lesions in the breast in relation to the pathomorphological verification.

The aim of this study was to evaluate the diagnostic value of the ultrasound B-mode imaging in the differentiation diagnostics of solid lesions in the breast in relation to the pathomorphological verification. From January to July 2010, 375 ultrasound breast examinations were conducted. The study enrolled 80 women aged 17-83, with 99 solid, focal lesions present in breasts, which were qualified for pathomorphological verification on the basis of the ultrasound examination. All patients underwent: the interview, physical examination, ultrasound examination and sonoelastography. The ultrasound features of the lesions, their vascularization patterns in the Doppler examination as well as the adjacent tissues were determined. Next, the focal lesions were categorized according to the BIRADS-US classification. The obtained results were analyzed statistically. In the group of 80 patients, 99 focal, solid lesions in breasts were visualized, including 39 neoplastic, malignant lesions (group I) and 60 lesions of benign nature (group II). The malignant lesions were often characterized by: greater size, irregular shape (34/39), prevalence of the anteroposterior dimension over the lateral-lateral dimension (22/39), acoustic shadowing (20/39), the margins not well-circumscribed (37/39), spiculated margins (16/39) and the presence of calcifications (14/39). The benign lesions were much more often hyper- and isoechogenic (14/60). In group I the lesions more often demonstrated the features of increased vascularization (29/39) and the presence of irregularly shaped vessels (23/29). This vascularization more often originated in the adjacent tissues. In the surroundings of the malignant neoplastic lesions, the presence of edema (16/39) and skin thickening (6/39) occurred more frequently and the abnormal axillary lymph nodes were more often diagnosed. The lesions of group I were assigned to the following BIRADS categories: BIRADS-US 4 (9 lesions) and BIRADS-US 5 (30 lesions). In group II, there was a prevalence of BIRADS-US 3 and 4 categories (58 lesions) and 2 lesions were classified to BIRADS-US 5 category. In the statistical analysis of the models based on BIRADS classification, it was demonstrated that BIRADS 4 showed the highest sum of sensitivity and specificity values of 173.6% in differentiation of the character of focal lesions in the breast (sensitivity 76.92%, specificity 96.67%).

The most common malignant neoplasm in Poland is breast cancer. Its incidence continues to increase. According to the latest statistical data, in 2009 more than 15 thousand new cases were reported in Poland(.

Currently, in the diagnostic process of the focal lesions in breasts the following examinations are conducted: palpation, imaging (i.e. mammography – MMG, ultrasound – US, magnetic resonance – MRI) as well as histopathological and/or cytological verification. A new technique of US imaging, which was introduced several years ago, is sonoelastography(. This method, by means of using elastic properties of tested tissues, allows for imaging of tissue deformability which is greater in lesions of benign nature and lesser in the majority of malignant neoplastic lesions due to the presence of desmoplastic growth of the connective tissue(. Therefore, sonoelastography constitutes a valuable supplementation of conventional ultrasound examination, enriching it with the biophysical features of the evaluated tissues. The sonoelastograms obtained, constitute a map of deformabilities of the examined tissues, which is placed on the classic B-mode image. In 2012, sonoelastography was introduced into Polish standard US breast examinations as an additional method, which aids in the differentiation diagnostics of focal lesions in breasts. It is also helpful in deciding whether the pathomorphological verification is essential(.

The studies, which have so far been carried out in the world and in Europe, confirm the usefulness of this method, particularly with respect to the evaluation of the lesions that are possibly benign in the B-mode tests (i.e. BIRADS 3) and those suspected of being malignant (i.e. BIRADS 4)(. As of today, there have been no such studies in Poland.

The article presented below is the first part of the publication whose aim was to determine the usefulness of sonoelastography in the differentiation of the character of solid lesions in the breast, as compared with the classic B-mode imaging, in relation to pathomorphological verification. The article has been prepared on the basis of the author's doctoral thesis entitled: The usefulness of sonoelastography in the diagnostics of solid lesions in the breast (written under the supervision of: Prof. Iwona Sudoł-Szopińska, defended on 25 October 2012 in the Cancer Centre, The Institute of Maria Skłodowska-Curie in Warsaw).

Material and methods

From January to July 2010 in the Ultrasonography Laboratory of the Cancer Centre – Institute in Warsaw, 375 US examinations were conducted due to focal lesions in breasts found by palpation, abnormalities determined in the imaging examinations or breast pain.

Out of this group, 80 women were selected for a sonoelastography examination. They were 17–83 years old with 99 solid, focal lesions present in breasts, which were qualified for histopathological and/or cytological verification on the basis of the classic breast ultrasound B-mode examinations.

All patients underwent: the interview, physical examination, imaging examinations: ultrasound examination and sonoelastography as well as histopathological and/or cytological verification.

The US examinations were conducted by means of Hitachi EUB-7500 HV apparatus with a linear transducer in the frequency range of 7.5–13 MHz with a compression sonoelastography option.

In accordance with the updated standards of the Polish Ultrasound Society (Polskie Towarzystwo Ultrasonograficzne, PTU) and ACR BIRADS-US lexicon(, the following ultrasound image features of the tested focal lesions were determined during sonomammography: echogenicity, echotexture, shape, orientation, boundaries, dimensions, vascularization, the presence of the acoustic shadowing, calcification, multifocality and multicentricity. The tissues surrounding the focal lesions were also evaluated for: edema, Cooper ligaments thickening, skin thickening as well as the presence of abnormalities in the axillary lymph nodes. The thickness ≥3 mm was considered a criterion for skin thickness. The edema was defined as the presence of hyperechogenic halo in the area of the focal lesion. The features of abnormal lymph nodes were: round shape, thickened hypoechogenic cortical zone or its eccentric thickening, uncircumscribed capsule of the node, atrophy or dislocation of its core zone (hilum).

When analyzing the two groups (I and II) with regard to the vascularization in the Doppler, power Doppler and color Doppler tests, 3 types of lesions were identified:

without visible vessels;

with few vessels (so-called hypovascular, 1–3 vessels in the lesion);

with vessels present (so-called hypervascular, more than 3 vessels in the lesion).

The distribution of vessels (peripheral, central vascularization or vessels going into the lesions from surrounding tissues) and their course (straight and irregular) were evaluated as well as the spectral analysis was carried out (resistive index – RI – measurement, using the cut-off value for benign lesions <0.85).

On the basis of the analysis of the ultrasound image features, the visualized focal lesions in breasts were assigned to the categories according to BIRADS-US classification.

The lesions were classified to individual categories: BIRADS-US 3, 4 and 5. The lesions of BIRADS-US 1 (physiological imaging) and BIRADS-US 2 (benign lesions) were not included in the study.

In all cases, the histopathological and/or cytological verification of the tested lesions was obtained.

The solid lesions categorized as BIRADS-US 3 were subject to cytological verification. The lesions categorized as BIRADS-US 4 were subject to cytological and/or histopathological verification. The lesions categorized as BIRADS-US 5 were subject to histopathological verification.

The patients with lesions categorized as BIRADS-US 3 and cytologically verified as benign underwent a follow-up US examination after 6 months.

The obtained results were analyzed statistically by means of statistical software packages (Statistica and IDAMS). The assumed statistical significance level was α≤ 0.05. Moreover, statistical models were created for individual BIRADS categories and the sensitivity, specificity as well as positive and negative predictive values were estimated. The threshold value for all categories of BIRADS classification was determined, which with the highest sum of sensitivity and specificity differentiated the character of the focal lesions in breasts.

Results

In the group of 80 patients, 99 focal, solid lesions in breasts were visualized. Thirty-nine lesions were of malignant character (group I of the lesions analyzed) and 60 lesions were of benign nature (group II of the lesions analyzed). The average age of patients in group I was 55.07 and it was significantly higher than the average age of the women in group II, which constituted 46.9 (p = 0.0067).

The most common carcinomas in group I were invasive ductal carcinomas (56.4%) and cribriform carcinomas (12.8%) (fig. 1). In group II the most common pathologies were fibrocystic breast changes (56.7%) and fibroadenomas (20%) (fig. 2).

Fig. 1

Results of pathological verification in patients from group I

Fig. 2

Results of pathological verification in patients from group II

In all 18 cases of the lesions categorized as BIRADS-US 3, fine-needle aspiration biopsy (FNAB) was performed, which confirmed the benign nature of the lesions. In the category BIRADS-US 4 (49 lesions), 27 FNAB and 31 histopathological verifications were conducted. In 9 cases both kinds of verifications were performed. In this group, 9 lesions were of neoplastic malignant character.

In the case of BIRADS-US 5 category, all lesions (total of 32) were subject to histopathological verification. Thirty instances were stated as malignant neoplasms in this group.

From among the analyzed US image features and elements of the examined focal lesions and surrounding tissues, which have been listed in material and methods, the features and elements which in a statistically significant manner differentiated malignant lesions from the benign ones were distinguished. The typical features of neoplastic, malignant lesions:

irregular shape (34/39);

prevalence of the anteroposterior dimension over the lateral-lateral dimension (22/39) and

the presence of acoustic shadow (20/39).

The benign lesions were much more often hyper- and isoechogenic (14/60) (tab. 1).

Tab. 1

The analyzed US features of the focal lesions in breasts in groups I and II

	Group	Lesion shape (irregular)	Lesion dimension (ant.-post.)	The presence of acoustic shadow	Hyper- and isoechogenic lesions
Number of lesions (% of the group)	I	34 (87,2%)	22 (56,4%)	20 (51,3%)	1 (2,6%)
	II	31 (51,7%)	17 (28,3%)	13 (21,7%)	14 (23,3%)
p		0,0003	0,0052	0,0023	0,0049
Sensitivity		87,17%	56,41%	51,28%	97,43%
Specificity		48,33%	71,66%	78,33%	23%
ppv		52,30%	56,41%	60,60%	45%
npv		85,29%	71,66%	78,33%	93,33%

ppv – is the proportion of persons with positive test results, in whom the examined condition or feature is present; npv – is the proportion of persons with negative test results, in whom the examined condition or feature is not present.

Out of all examined features, the echogenicity was a feature which differentiated the focal lesions with the highest sensitivity. It constituted 97.4% for lesions of benign character.

In group I, the following lesions were detected significantly more often:

lesions of greater dimensions;

lesions of margins which are not well-circumscribed (37/39) and spiculated (16/39);

lesions with calcifications (14/39) (tab. 2).

Tab. 2

The analyzed US features of the focal lesions in breasts in groups I and II

	Group	Dimensions (average maximum values) [mm]	Not well-circumscribed margins	Spiculated margins	Calcifications
Number of lesions (% of the group)	I	16,71	37 (94,9%)	16 (41%)	14 (35,9%)
	II	12,43	39 (65%)	3 (5%)	8 (13,3%)
p		0,0022	0,0006	0,0074	0,0083
Sensitivity		-	94,87%	41,02%	35,89%
Specificity		-	35%	95%	86,66%
ppv		-	48,68%	84,21%	63,63%
npv		-	31,3%	71,25%	67,53%

ppv – is the proportion of persons with positive test results, in whom the examined condition or feature is present; npv – is the proportion of persons with negative test results, in whom the examined condition or feature is not present.

Solid, isoechogenic lesion, category: BIRADS-US 3 (histopathological verification – fibroadenoma)

Solid, hypoechogenic lesion with the prevalence of the anteroposterior dimension over the lateral-lateral dimension, category: BIRADS-US 4 (histopathological verification – preinvasive ductal carcinoma)

Solid hypoechogenic lesion of irregular shape with the presence of calcifications, the margins are not well-circumscribed from the surroundings, acoustic shadow behind the lesion, BIRADS-US 5 (histopathological verification – invasive ductal carcinoma)

The features which differentiated the benign lesions from the neoplastic malignant ones best (i.e. with the highest sensitivity and specificity of 94.9% and 95% respectively) were the spiculated and not wellcircumscribed margins of the focal lesions.

The majority of neoplastic, malignant lesions (29/39) presented the feature of increased vascularization. The irregular course of vessels was detected significantly more often in group I (23/29). This vascularization much more often originated in the adjacent tissues (tab. 3).

Tab. 3

The vascularization of the focal lesions in breasts in groups I and II

	Group	Increased vascularization	Vascularization originating in adjacent tissues	Vessels of irregular course
Number of lesions (% of the group)	I	29 (74,4%)	19 (65,5%)	23 (79,3%)
	II	21 (35%)	4 (19,1%)	2 (9,5%)
p		0,0000	0,0011	0,0000
Sensitivity		37,93%	65,51%	79,31%
Specificity		90,47%	80,95%	90,47%
ppv		84,61%	82,6%	92%
npv		51,35%	62,96%	76%

ppv – is the proportion of persons with positive test results, in whom the examined condition or feature is present; npv – is the proportion of persons with negative test results, in whom the examined condition or feature is not present.

Solid, hypoechogenic lesion with spiculated margins, irregular shape and edema in the adjacent tissues, BIRADSUS 5 (histopathological verification – invasive ductal carcinoma)

Out of the vascularization types, the character of the focal lesions was best differentiated by the presence of the vessels with irregular course (specificity 90.5%, ppv 92%).

In the surroundings of the malignant neoplastic lesions, the presence of edema (16/39), skin thickening (6/39) and abnormal axillary lymph nodes were significantly more frequently observed (tab. 4). Each listed abnormality, apart from Cooper ligaments thickening, significantly differentiated the neoplastic malignant lesions from the benign ones.

Tab. 4

US features of the tissues which surround the examined focal lesions and the presence of abnormal axillary lymph nodes

	Group	Edema	Skin thickening	Abnormal axillary lymph nodes
Number of lesions (% of the group)	I	16 (41%)	6 (20%)	14 (46,7%)
	II	13 (21,7%)	1 (2%)	1 (2%)
p		0,0386	0,0107	0,0045
Sensitivity		41,02%	20%	46,66%
Specificity		78,33%	97,95%	97,95%
ppv		55,17%	85,71%	93,33%
npv		67,14%	66,66%	75%

ppv – is the proportion of persons with positive test results, in whom the examined condition or feature is present; npv – is the proportion of persons with negative test results, in whom the examined condition or feature is not present.

Solid hypoechogenic lesion with increased vascularization and the presence of vessels with irregular course in the area of the lesion and in the adjacent tissues, BIRADS-US 5 (histopathological verification – invasive ductal carcinoma)

Abnormal lymph nodes were diagnosed in 14 patients in group I (14/30) and in 1 patient in group II (1/49). This last feature proved the best in differentiating the focal lesion character in breasts (specificity and ppv were 98% and 93.3% respectively).

The neoplastic malignant lesions were assigned to the following BIRADS categories: BIRADS-US 4 (9 lesions) and BIRADS-US 5 (30 lesions). No lesion was assigned to category BIRADS-US 3.

Skin thickening (the crosses) up to 5 mm in a patient with invasive ductal carcinoma

Abnormal, enlarged axillary lymph node in a patient treated for ductal breast carcinoma: hypoechogenic, without a visible sinus and of an increased vascularization in the area of the capsule and hilum of the node

The group of benign lesions was dominated by the lesions classified as BIRADS-US 3 and 4 (58 lesions). Merely 2 lesions were classified to BIRADS-US 5 category.

The data included in tab. 5 were drawn up by means of 2 statistical models based on the discriminant analysis: BIRADS 3/4 and BIRADS 4/5. In each of these models, the indicators of sensitivity, specificity, ppv and npv were compared in order to identify the threshold value, which would be the best means to aid in the differentiation of benign lesions from the malignant ones.

Tab. 5

The percentage distribution of focal lesions in breasts according to BIRADS-US classification

	Group	BIRADS 3	BIRADS 4	BIRADS 5	Total	%
Number	Group I	0	9	30	39	39,39
% of the column		0,00	18,37	93,75
% of the verse		0,00	23,08	76,92
Number	Group II	18	40	2	60	60,61
% of the column		100,00	81,63	6,25
% of the verse		30,00	66,66	3,44
Total		18	49	32	99
%		18,18	49,49	32,32		100,00

In the statistical analysis of the models based on BIRADS classification, it was demonstrated that BIRADS-US 4 showed the highest sum of sensitivity and specificity values of 173.6% (with the sensitivity of 76.92% and the specificity of 96.67%). This model proved to be the best in differentiating lesions in both groups.

Discussion

The BIRADS system was introduced into mammography breast examinations in 1993. It standardized the descriptions of these examinations by issuing the lexicon of pathological lesions(. Numerous studies published in subsequent years proved its usefulness in estimating the risk of focal lesion malignancies in breasts(. For this reason, after 10 years, the uniform BIRADS classification system was introduced into US examinations. Currently, it is postulated that the BIRADS lexicon should be supplemented with the evaluation of lesions by means of sonoelastography(.

Ultrasonography is particularly recommended for women with a prevalence of glandular structure of breasts. The latest publications indicate its greater usefulness, as compared to MMG, in differentiating the breast lesion character in this group of patients(.

The breast cancer incidence increases after the age of 30 reaching its peak at the age of 50 and remaining on a high level in subsequent years(. In this study, the average age in a group of 30 women with diagnosed malignant neoplastic lesions constituted 55.07 and was significantly higher than in the group of 49 patients with benign lesions, which was 46.9. Itoh et al.(, in their study, which enrolled 111 Japanese patients, observed a similar phenomenon. The average age of patients with diagnosed breast cancer constituted 52.9 and the age of patients with benign lesions was 47.4. German authors, Schaefer et al.(, on the other hand, in a group of 193 examined patients, noted a much greater age average for women with diagnosed breast cancer, which constituted 62.4, as compared with 49.3 for patients with benign lesions.

An important parameter form the point of view of therapeutic procedure is the dimension of the examined focal lesion in breasts. In this study, on the basis of the analysis of the average lesion dimensions in both groups, it was stated that the neoplastic malignant lesions were significantly larger than the benign ones. The average value of the largest neoplastic malignancies constituted 16.1 mm and in the case of benign lesions, the value amounted to 12.4. Similar differences in the size of lesions were obtained by Schaefer et al.( and Itoh at al.( However, no significant differences in the dimensions between the malignancies and benign lesions were stated in the research of Wojcinski et al.(

Out of the diagnosed malignant neoplasms, the most frequent condition was invasive ductal carcinoma (56.4% of malignancies) and in the group of benign lesions the fibrocystic breast lesions (56.7% of benign changes). In the research of the Japanese scholars(, in which the age average was similar to the one in the author's own study, the ductal carcinoma constituted 80.7% and the most frequently occurring benign lesions constituted the so-called ANDI (aberrations of normal development and involution), an equivalent of fibrocystic breast lesions, which constituted 40.7% of the cases.

As far as echogenicity is concerned, the benign lesions were significantly more frequently hyperechogenic and isoechogenic than the malignant ones. The neoplastic malignant lesions significantly more often presented: acoustic shadow, irregular shape, greater size, prevalence of the anteroposterior dimension over the lateral-lateral dimension, calcifications, increased vascularization with irregular course, edema in adjacent tissues, skin thickening and abnormal axillary lymph nodes. Similar analyses were conducted by other authors. On the basis of the BIRADS lexicon, Hong et al.( analyzed the ultrasound features of 403 focal lesions in breasts and evaluated their positive and negative predictive values for benign and malignant lesions. The high ppv value was obtained for the following features typical of malignant neoplastic lesions: spiculated margins (86%), irregular shape (62%) and the prevalence of the anteroposterior dimension over the lateral-lateral dimension (69%). In the aforementioned study, 76% of the lesions with the presence of calcifications were of malignant character and the acoustic shadow behind the lesion, which was observed in 49% of cases, constituted a statistically significant way to differentiate malignant lesions from the benign ones. With an exception of the calcifications, these results coincide with the ones presented herein (respective results in the author's own study: 84.2%, 52.3%, 56.4%, 35.9% and 51.3%).

For benign lesions, however, the researchers stated high npv values for the following features: well-circumscribed boundaries (90%), oval shape (84%) and the prevalence of the lateral-lateral dimension over the anteroposterior dimension (78%). These results, except for the well-circumscribed boundaries for which npv in the author's own study was 31.3%, were similar to the researcher's own results and constituted 85.3% and 71.7% respectively.

Stavros et al.( compared the sensitivity of individual elements of US image and a set of ultrasound features of 750 focal lesions in breasts. The individual features which indicated the malignant character of lesions with the highest sensitivity were: angular margins (83%), slight bulges on the outlines (75%) as well as low echogenicity (hypoechogenic lesions – 69%). In the group of evaluated lesions, the similarity of the results with the outcomes of the author's own study concerned the outlines. The obtained level of sensitivity was higher – 94.4%. Upon total analysis of all 11 examined US features, the sensitivity of US examination in the authors’ research amounted to 98.4%. With respect to benign lesions, the high level of npv obtained by the authors concerned hyperechogenic lesions (100%), which was consistent with the observations of the author of this article, where the benign lesions were hyperechogenic and/or isoechogenic (npv of 93.3%).

Another analyzed parameter, which, in the study of the author of this article, significantly differentiated the character of focal lesions in breasts was their vascularization. In the literature, there are no explicit data concerning the usefulness of this parameter due to high heterogeneity of breast carcinomas, the influence of the menstrual cycle on the flow values as well as the influence of the administered hormonal therapy on the flow in the vessels in the area of the mammary gland. It is known, however, that neoplastic lesions, particularly those highly malignant, stimulate the process of neovascularization and the newly formed vessels are usually localized on the periphery of the lesion and have a winding course as well as sinusoidal arteriovenous connections(. In the material of the author of this article, the features of increased vascularization were prevalent as far as neoplastic malignant lesions are concerned. The vessels presented a winding course and permeated to the focal lesions from the adjacent tissues. As for the benign lesions, only single vessels with straight course were visible. In the author's own research, the RI indicator did not differentiate the focal lesion character in a significant manner. The Doppler examinations of 826 focal lesions in breasts conducted by del Cura et al.(, showed a significantly larger number of vessels in malignant tumors than in benign ones, which was confirmed in the study of the author of this article. In a spectral test the value of the RI indicator (on average 0.82) was higher for malignant lesions. This, however, was not confirmed in the author's own research. In addition, Algül et al.( on the example of 38 focal lesions in breasts also proved that RI indicator of value above 0.8 differentiates the character of focal lesions in breasts in a statistically significant way. The remaining observations of the authors connected with the vascularization of the focal lesions confirmed the author's own studies.

As for the changes in the tissues surrounding the examined focal lesions in breasts, such as Cooper ligament thickening, skin thickening, edema or structure disorders, the results obtained were similar to those obtained by Hong et al.( The authors found abnormalities in the tissues around the focal lesions in 14% of the cases including as much as 75% of neoplastic, malignant lesions. In the research of the author of this article, edema in the surroundings of the neoplastic focal lesions occurred in 41% of cases and skin thickening – in 20% of neoplastic malignant lesions. Thus, occurrence of the abnormalities in the surroundings of the focal lesions proved to be an additional feature which significantly differentiated the character of the focal lesions in the breast.

An integral part of US examinations of focal lesions in breasts is the evaluation of axillary lymph nodes. The presence of pathological nodes in this site constitutes an essential prognostic factor and affects the manner of local treatment of breast cancer. In the author's own material the abnormal lymph nodes were visualized in 47% of patients with neoplastic malignancies and in 2% of patients with benign lesions. In only one patient with a benign lesion, a node showed US features typical of metastatic condition, but in a pathomorphological verification, no neoplastic cells were found. In a group of 425 patients with invasive breast carcinoma, Choi et al.( detected abnormal lymph nodes in 74% of cases – the sensitivity constituted 88.6%, npv 83%, specificity merely 36% and ppv 47%. Similarly to the material of the author of this article, the researchers thoroughly analyzed ultrasound features which allow for the differentiation between malignant and benign lesions and they showed that the thickening of the cortical zone above 3 mm constitutes the most reliable predictive feature for the diagnosis of metastatic lesions in lymph nodes. In their study, a much higher specificity was obtained – 97.8% as well as a higher positive predictive value – 93.3%.

On the basis of morphological features of focal lesions in breasts, which were analyzed during US exam, BIRADS categories were assigned. In the author's own study, all neoplastic malignant lesions were classified as BIRADS 4 (23.1%) and BIRADS 5 (76.9%). The lesions of benign character, on the other hand, were the most frequently assigned to category BIRADS 4 (66.7%). BIRADS 5 category two lesions were classified (3.3%) and the remaining ones were categorized as BIRADS 3 (30%). One of two false positive lesions was duct papilloma with fibrosis and sclerosing, which was eventually treated surgically. The other lesion was mastitis, which regressed upon treatment.

Schaefer et al.( in their prospective study based on 193 focal lesions in breasts, demonstrated an arrangement of benign lesions into individual BIRADS categories, which was similar to the one presented by the author of this study. Out of 129 benign lesions, 72.1% were classified as BIRADS 4a and 4b and 3.9% were given BIRADS 5 category. The remaining benign lesions were described as BIRADS 2 (2.3%) and BIRADS 3 (21.7). In the author's study, BIRADS 2 categories were not considered since they do not require pathological verification which constituted a point of reference for the results of ultrasonography and sonoelastography examinations. On the other hand, in the group of malignant lesions, the authors assigned the changes to BIRADS 4 and BIRADS 5 categories with 42.2% and 57.8% respectively. In the aforementioned research, the percentage of lesions classified as BIRADS 4 was nearly twice as large as in the author's own material and BIRADS 5 was nearly 20% smaller. The reasons for such discrepancies in lesion classification to BIRADS 4 or 5 (particularly in 4b and 5) are not clear. The greater number of neoplastic lesions assigned to BIRADS 5 in the author's own material, undoubtedly resulted from the specific nature of the oncology institute where the examinations recommended to the patients with breast cancer are conducted. Other reasons, such as errors in lesion evaluation, should be excluded since the generally accepted the ACR BIRADS lexicon constituted the basis for evaluation.

For the purpose of indicating the threshold value of the highest sensitivity and specificity in character differentiation of focal lesions in breasts, individual models drawn up on the basis of BIRADS were evaluated. BIRADS 4/5 occurred to be the best model. Here, the sum of sensitivity and specificity amounted to 173.6% (sensitivity 76.9%, specificity 96.67%). Based on this model, only 2 cases of false negative lesions and 9 false positive ones were described. In the research quoted above, i.e. of Schaefer et al.(, the model with the highest threshold value was also BIRADS 4/5. Here, however, the sensitivity was lower (57.8%) but specificity was similar (96.1%) (the sum – 153.9%), as compared with the author's own study. The lower sensitivity in the group of German researchers for this cut-off value resulted from increased percentage of false negative results. High indicators of specificity in the author's own research and in the study of German researchers attests to the low percentage of false negative results.

The results of the B-mode imaging in the differentiation diagnostics of solid, focal lesions in breasts in relation to the pathomorphological verification, which were presented and discussed above, confirm high usefulness of this exam in estimating the character of solid, focal lesions in breasts with the application of the threshold value of BIRADS 4/5.

In the second part of the article (ed. note – which will be published in the 52nd issue of “Journal of Ultrasonography”) the above discussed results will be discussed with respect to sonoelastography.

Tab. 6

The values of sensitivity, specificity, ppv and npv indicators and the sum of sensitivity and specificity for individual models of BIRADS-US classification

Statistical models based on BIRADS	BIRADS 3/4		BIRADS 4/5
Evaluation	Malignant	Benign	Malignant	Benign
Malignant lesion	39	0	30	9
Benign lesion	42	18	2	58
Sensitivity	100,00		61,7–87,4	76,92
Specificity	30,00		88,6–99,1	96,67
ppv	48,15		79,9–98,3	93,75
npv	100,00		76,4–92,8	86,57
Sensitivity + specificity	130,00			173,59

ppv – is the proportion of persons with positive test results, in whom the examined condition or feature is present; npv – is the proportion of persons with negative test results, in whom the examined condition or feature is not present.