The value of dynamic contrast-enhanced magnetic resonance imaging combined with apparent diffusion coefficient in the differentiation of benign and malignant diseases of the breast

Abstract

Background

The value of combined dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) and apparent diffusion coefficient (ADC) histogram analysis for the diagnosis of breast cancer has not been evaluated in previous studies.

Purpose

To investigate the diagnostic value of DCE-MRI combined with ADC in benign and malignant breast lesions.

Material and Methods

The clinicopathological imaging data included 168 patients (177 lesions) with breast lesions who underwent convention breast MRI, DCE-MRI, and diffusion-weighted imaging (DWI); they were divided into the benign lesion group (n = 39) and malignant lesion group (n = 129) based on pathology.

Results

Using the type III outflow curve as a diagnostic criterion for malignant breast lesions, the diagnostic sensitivity was 76.9%, the specificity was 80%, the correct rate was 72.2%, and its area under the curve (AUC) was 0.823. Using an enhancement ratio > 100% as a diagnostic criterion for malignant breast lesions, the sensitivity was 61.5%, specificity was 80%, and AUC was 0.723. Using > 3 ipsilateral vessels as a diagnostic criterion for malignant lesions in the breast resulted in a diagnostic sensitivity of 81.6%, a specificity of 80.8%, and an AUC of 0.805.

Conclusion

The type of time intensity curve DCE-MRI, the early enhancement rate in the first phase, the number of ipsilateral vessels, and the ADC full volume histogram of the blood supply score and DWI are valuable in the diagnosis of benign and malignant breast lesions.

Keywords

Breast disease dynamic enhancement diffusion-weighted imaging magnetic resonance imaging

Introduction

Breast magnetic resonance imaging (MRI) has a wide range of clinical applications by way of good soft-tissue resolution and no radiation. For the diagnosis of breast diseases by MRI in clinical work, we mainly refer to the Breast Imaging Reporting and Data System (BI-RADS) from the American College of Radiology. The system mainly evaluates the benign and malignant nature of the lesions according to the morphological characteristics, and has certain subjectivity, although the diagnostic sensitivity is high and the specificity is low (1). Dynamic contrast-enhanced MRI (DCE-MRI) of the breast provides high-resolution morphological information as well as functional information of neovascularization (2) and has good sensitivity and specificity for breast lesions. Allarakha et al. (3) suggested that DCE-MRI has important value in the diagnosis and prognosis evaluation of breast cancer. Braman et al. (4) also agreed that DCE-MRI has excellent sensitivity and specificity for the diagnosis of breast cancer.

MR diffusion-weighted imaging (DWI) is an MRI functional imaging that non-invasively detects the diffusion movement of water molecules within living tissues, and the automatically generated apparent diffusion coefficient (ADC) contains abundant texture information to quantitatively detect the extent of restricted diffusion of water molecules (5). Previous studies mostly used the method of delineating a single-slice region of interest (ROI) to obtain the mean ADC value to quantify the restricted diffusion of water molecules at the lesion site (6,7). However, the mean ADC value cannot comprehensively reflect the overall ADC information of the lesion, and it ignored the heterogeneity within benign and malignant lesions (8). ADC histogram analysis is a method used to scan the ROI layer by layer and then generate three-dimensional (3D) images. It uses the information of the ADC value of the whole lesion to describe the degree of heterogeneity limited by diffusion of water molecules at the lesion site and has shown superiority in tumor characterization, grading, and predicting patient survival (9,10). In addition, the histogram analysis of ADC can provide a more objective and accurate basis for the differential diagnosis of benign and malignant breast lesions (11,12). Furthermore, researchers have combined DCE-MRI and DWI to predict breast malignancy, which can improve differentiating performance (13).

The value of combined DCE-MRI and ADC for the diagnosis of breast cancer has not been well evaluated in previous studies. The aim of the present study was to combine each parameter of DCE-MRI and ADC and analyze their sensitivity and specificity for the diagnosis of benign and malignant breast lesions, and to investigate their value in the differential diagnosis of benign and malignant breast lesions.

Material and Methods

Patients

A total of 168 patients (177 lesions) with breast lesions diagnosed by pathology after breast MRI examination in the medical imaging center of Renmin Hospital from January 2018 to October 2020 were retrospectively analyzed. All patients were divided into the benign lesion group (n = 39) or malignant lesion group (n = 129) based on pathology.

Inclusion criteria were as follows: (i) before breast MRI examination, no invasive examination or chemoradiotherapy was performed in both breasts; (ii) pathological results of breast lesions were obtained by surgery or biopsy after MRI examination; (iii) the image was clear enough to meet the requirements of delineating ROI; and (iv) the scan image was of good quality without obvious motion artifacts.

Exclusion criteria were as follows: (i) patients who had undergone puncture, neoadjuvant chemotherapy, radiotherapy, or surgery; (ii) the quality of the MRI scan was poor, and the lesions could not be judged; and (iii) the ADC value or dynamic enhancement curve could not be measured and could not be used for analysis.

This study was approved by the ethics committee of Renmin Hospital of Hubei University of Medicine (No. syrmyy2018-023), and all patients signed the informed consent.

Imaging methods

The patients were scanned in the prone position with the bilateral mammary glands naturally hanging inside the mammary dedicated coils using a Siemens Magnetom Skyra 3.0 T superconducting MR scanner with an eight-channel mammary-dedicated MR coil. The scanning sequence and related parameters were as follows: (i) conventional MRI scan: axial: T1-weighted (T1W) image unenhanced stamper sequence (TR = 5.7 ms, TE = 2.4 ms, matrix = 358 × 448, bandwidth = 220, slice thickness = 4 mm, number of excitations [NEX] = 1) and T2-weighted (T2W) image stamper sequence (TR = 7900 ms, TE = 75 ms, matrix = 358 × 448, bandwidth = 252, slice thickness = 4 mm, NEX = 2) and sagittal T2 stamper sequence (TR = 3790 ms, TE = 73 ms, matrix = 192 × 256, bandwidth = 260, slice thickness = 4 mm, NEX = 1). (ii) DWI scan: axial: scan level co-axial axial piezoelectric sequence using a single excitation planar echo imaging (EPI) sequence with b-values of 0 and 800 s/mm². (iii) DCE scanning: enhancement was preceded by a masked scan, and after confirming the scan range, gadoteric acid was administered intravenously through the elbow using a high-pressure syringe at 2.0 mL/s with a dose of 0.1 mmol/kg body weight, after which the tube was flushed with 20 mL normal saline. The scanning was performed by 3D cross-sectional breast volume without interval scanning technique with the following parameters: TR = 4.5 ms; TE = 1.7 ms; bandwidth = 410; slice thickness = 1.5 mm; scan time = 40–60 s; one-phase plain scan; and five-phase enhancement.

Image analysis and data measurement

Images obtained from MR scans were passed to a Siemens syngo Workplace postprocessing workstation for analytical measurement of lesion characteristics by two experienced imaging diagnosticians without known pathology, resolved by consensus when controversial. The maximum slice of the lesion was selected to outline the ROI layer by layer along the edge of the lesion, avoiding macroscopically visible areas of hemorrhage, liquefaction, necrosis, and cystic change, draw a time-signal intensity curve (TIC), and calculate the early enhancement rate according to the TIC. The TIC was drawn and analyzed by two radiologists with > 10 years of experience in the diagnosis of breast MRI. Each measurement was repeated three times to take the average value. When the two opinions were inconsistent, the final diagnosis was made by referring to the third physician (who had been engaged in imaging diagnosis for > 15 years) for decision negotiation. TIC classification is as follows (14): type I = persistent (persistent enhancement of ≥10% initial enhancement); type II = plateau (constant signal intensity after peak attainment of ± 10% initial enhancement); or type III = clearance (decrease in signal intensity after peak enhancement > 10% initial enhancement). The early enhancement rate of the first phase can be shown as (Ie-Ipre)/Ipre × 100%, where Ie the signal value of the first sequence after enhancement and Ipre is the signal value before enhancement.

The resultant maximum intensity projection (MIP) image from the first sequence subtraction after the enhancement scan was selected for the analysis of the mammary vasculature. The number of vessels ≥3 cm in length or ≥2 mm in greatest transverse diameter within the ipsilateral mammary gland was recorded; in disagreement, a conclusion was reached by consensus. Using the blood supply scoring method proposed by Sardanelli et al. (15), the four grades of poor or no blood supply to the breast (score 0), low blood supply (score 1), moderate blood supply (score 2), and high blood supply (score 3) were assigned, and the number of representative vessels corresponding to them was 0, 1, 2–4, and ≥ 5. A score of 2–3 was considered to be increased blood supply, which was diagnosed by MRI as a predisposing malignant lesion.

ADC maps were automatically generated after DWI sequence scans were completed, exported in Bitmap (.bmp) format, and all images remained consistently windowed. Histogram analysis was performed on the full-volume ADC maps: combined with the enhanced images, the whole layer where the lesion was located within the ADC map was selected, imported into MaZda software layer by layer to outline the ROI along the edge of the lesion, and the software automatically generated gray-scale histograms of the ROI and obtained histogram parameters (mean, variance, skewness, kurtosis and the 1st, 10th, 50th, 90th, and 99th percentiles, etc.).

Statistics

Statistics were performed using SPSS version 22.0 software, normally distributed metrics were presented as mean ± standard deviation, and comparisons between the two groups were performed using the independent samples t test. The chi-square test was used for comparison between count data; receiver operating characteristic (ROC) curves were used to evaluate the diagnostic value of each parameter for benign and malignant breast lesions. The sensitivity, specificity, concordance rate, Youden index, and likelihood ratio of MRI for the diagnosis of malignant breast lesions were calculated. When the AUC was in the range of 0.5–1.0, it indicated that the full-volume ADC histogram parameters had high diagnostic trueness for benign and malignant breast lesions. The inspection level was set as α = 0.05. The logistic assessment was used for each parameter combined. P < 0.05 indicated statistically significant differences.

Results

Clinical and pathological features

A total of 168 patients (all women) with 177 lesions were included in the study, including 39 patients (43 lesions, mean age = 35.1 ± 7.8 years) in the benign lesion group and 129 patients (134 lesions, mean age = 52.2 ± 10.9 years) in the malignant lesion group. The age of onset of malignant lesions was higher than that of benign lesions, and the difference was statistically significant (P < 0.05). The mean size of the benign tumors was 1.68 ± 0.35 cm, and that of the malignant tumors was about 1.77 ± 0.41 cm, with no statistically significant difference (P = 0.016).

The 43 benign lesions included 29 fibroadenomas, three adenopathies, six mastitis, and five intraductal papillomas. The 134 malignant lesions included 72 invasive breast carcinomas, 16 co-existing invasive breast carcinomas and ductal carcinomas in situ, 33 ductal carcinomas in situ, two invasive lobular carcinomas, five mucinous carcinomas, one adenoid cystic carcinomas, four intraductal papillary carcinomas, and one breast metastasis.

Imaging characteristics and type of TIC of benign and malignant breast lesions

There are distinct differences between benign and malignant breast lesions in terms of imaging characteristics. Benign lesions were mostly round or doughnut-shaped with well-defined borders, uniform signal, and unenhanced or mildly to moderately homogeneously enhancing on contrast-enhanced scans. Malignant lesions mostly showed irregular lesion morphology, lobulated margins, visible spiculation sign, indistinct borders, less homogeneous signal, and markedly uneven enhancement in contrast-enhanced scans. Among 43 benign lesions, the time signal curves showed 33 lesions of type I (33/43, 76.7%), eight lesions of type II (8/43, 18.6%), and two lesions of type III (2/43, 4.65%). Among the 134 malignant lesions, the time signal curves showed four lesions of type I (4/134, 2.98%), 25 lesions of type II (25/134, 18.6%), and 105 lesions of type III (105/134, 78.3%). Typical benign lesion features of the breast are shown in Fig. 1, and typical malignant lesion features of the breast were displayed in Fig. 2. Using the type III outflow curve as a diagnostic criterion for malignant lesions of the breast, the diagnostic sensitivity was 76.9%, specificity 80%, trueness 72.2%, and area under the curve (AUC) was 0.823.

Fig. 1.

Typical benign lesion features of the breast in a 36-year-old female patient. (a) Axial T2WI-FS with a round-like long T2 hyperintense shadowing and well-defined lesion borders. (b) DCE-MRI stage I: the lesion is seen as moderately homogeneously enhancing. (c) TIC, showing a type II (plateau type) with a rapid rate of intensification in phase 1. (d) MIP obtained from the first sequence subtraction after the enhancement scan, suggesting that the number of vessels with a maximum transverse diameter of ≥ 2 mm within the ipsilateral mammary gland was 1 (low blood supply), and the Sardanelli blood supply score was 1 point. (e) Original DWI map showing the lesion as hyperintense. (f) ADC raw map showing the lesion as slightly hypointense with a measured ADC value of 1054.5 ± 157.5 × 10⁻³mm²/s. (g) Lesion full-volume ADC histogram shown by outlining an ROI (blue section) along the edge of the lesion after the ADC map was imported into MaZda software. (h) Lesion ADC histogram parameter value. (i) Pathological gross specimen after surgical resection. (j) Pathological section characteristic, showing as fibroadenoma; immunohistochemistry results were: ER (+), CK5/6 (+), P63 (+), SMA (+), Ki67 (low proliferation). ADC, apparent diffusion coefficient; DCE-MRI, dynamic contrast-enhanced magnetic resonance imaging; DWI, diffusion-weighted imaging; ER, estrogen receptor; MIP, maximum intensity projection; ROI, region of interest; SMA, smooth muscle actin; T2WI-FS, T2-weighted image fat suppression; TIC, time-intensity curve.

Fig. 2.

Typical malignant lesion features of the breast in a 51-year-old female patient. (a) Axial T2WI-FS, with irregular clumps of slightly elongated T2 signal shadowing and multiple lobulations of the lesion. (b) DCE-MRI stage I: the lesion has markedly heterogeneous enhancement with spiculation at the margins. (c) TIC, showing type III (outflow type) with a rapid rate of intensification in phase 1. (d) MIP obtained from the first serial subtraction after the enhanced scan, which suggested that the blood vessels in the right breast mass were significantly thickened, increased, and tortuous, with an abundant blood supply. The number of vessels ≥ 3 cm in length or ≥ 2 mm in maximum transverse diameter within the ipsilateral mammary gland was ≥ 5 (high blood supply), and the Sardanelli blood supply score was 3 points. (e) Original DWI map showing the lesion as hyperintense. (f) ADC raw map showing the lesion as slightly hypointense with a measured ADC value of 731.63 ± 75.18 × 10⁻³mm²/S. (g) Lesion full-volume ADC histogram shown by the ROI (blue section) outlined along the lesion margin after importing the ADC map into MaZda software. (h) Lesion ADC histogram parameter value. (i) Pathological gross specimen after surgical resection. (j) Pathological section characteristic, showing as invasive ductal carcinoma, immunohistochemistry results showed: CD34 (vascular +), CK5/6 (−), CK7 (+), E-cad (cell membrane +), ER (−), GATA-3 (+), HER-2 (3 +), Ki67 (Li40%), P120 (cell membrane +), P53 (mutant +), P63 (−), PR (−). ADC, apparent diffusion coefficient; DCE-MRI, dynamic contrast-enhanced magnetic resonance imaging; DWI, diffusion-weighted imaging; ER, estrogen receptor; MIP, maximum intensity projection; ROI, region of interest; SMA, smooth muscle actin; T2WI-FS, T2-weighted image fat suppression; TIC, time-intensity curve.

Malignant lesions showed a more rapid enhancement rate of DCE-MRI stage I than benign lesions

The enhancement rate of DCE-MRI stage I was described as the relative strengthening degree of the lesion at the early stage of enhancement and was categorized as slow (<50%), moderate (50%–100%), and rapid (>100%). There were significant differences (P < 0.05) in the rate of early intensification of stage I between benign and malignant breast lesions (Table 1). The rate of early intensification in stage I of malignant lesions was mainly rapid (enhancement rate > 100%). Its sensitivity, specificity, and AUC for the diagnosis of malignant lesions of the breast were 61.5%, 80%, and 0.723, respectively.

Table 1.

Enhancement rate of DCE-MRI stage 1 for benign and malignant breast lesions.

Category	Slow (< 50%)	Moderate (50%–100%)	Rapid (> 100%)	χ²	P
Benign	12	19	12	20.688	0.000 < 0.05
Malignant	12	34	88

DCE-MRI, dynamic contrast-enhanced-magnetic resonance imaging.

The number of ipsilateral vessels and blood supply score of malignant lesions were higher than those of benign lesions

The number of ipsilateral vessels and blood supply score of malignant breast lesions were significantly higher than those of benign breast lesions (P < 0.05) (Table 2). The sensitivity, specificity, and AUC of malignant breast lesions were 81.6%, 80.8%, and 0.805, respectively.

Table 2.

Comparison of ipsilateral vascular number and blood supply score between benign and malignant breast lesions.

Category	Benign	Malignant	t	P
No. of blood vessels	1.73 ± 0.98	4.59 ± 1.27	–13.522	0.026 < 0.05
Blood supply score	1.56 ± 0.51	2.71 ± 0.67	–10.328	0.020 < 0.05

Evaluation of the effect of ipsilateral breast blood supply increase in the diagnosis of breast malignant lesions

The sensitivity was 98.5%, specificity was 81.3%, and correct rate was 94.3% for the diagnosis of malignant lesions of the breast when the ipsilateral blood supply was increased, Youden index J was 0.798, the positive likelihood ratio LR (+) was 5.26, and the negative likelihood ratio LR (–) was 0.01. The sensitivity, specificity, and correct rate of the diagnosis of malignant lesions of the breast for this criterion were high; however, the Youden index and LR (+) values were not high (Table 3).

Table 3.

Results of malignant lesions diagnosed by increased blood supply to the ipsilateral breast.

Pathologic diagnosis	Ipsilateral breast blood supply increase		Total
Pathologic diagnosis	Benign	Malignant	Total
Benign	132	2	134
Malignant	8	35	43
Total	140	37	177
Sensitivity (%)	98.5
Specificity (%)	81.3
Correct rate (%)	94.3
Youden index J	0.798
Positive likelihood ratio LR (+)	5.26
Negative likelihood ratio LR (−)	0.01

Comparison of full volume ADC histogram parameters between benign and malignant breast lesions

In the ADC histogram parameters of the malignant lesions group, the mean, variance, 1st, 10th, 50th, 90th, and 99th quantiles were significantly lower than those of benign lesions group (P < 0.05), the skewness was higher than that of benign lesions group (P < 0.05), and there was no significant difference in kurtosis between the malignant and benign lesion groups (P > 0.05) (Table 4).

Table 4.

Comparison of full volume ADC histogram parameters between benign and malignant breast lesions.

Parameters	Benign	Malignant	t	P
Mean	158.95 ± 33.06	118.26 ± 16.49	10.720	0.000 < 0.05
Variance	771.02 ± 331.90	660.65 ± 173.02	2.839	0.001 < 0.05
Skewness	–0.35 ± 0.85	0.26 ± 0.65	–4.326	0.000 < 0.05
Kurtosis	0.31 ± 1.28	0.37 ± 1.25	–0.272	0.786
1st percentile	94.05 ± 22.83	64.78 ± 12.23	10.808	0.000 < 0.05
10th percentile	124.76 ± 33.64	88.34 ± 16.22	9.569	0.000 < 0.05
50th percentile	159.35 ± 32.79	118.21 ± 16.88	7.898	0.000 < 0.05
90th percentile	190.84 ± 31.68	152.36 ± 22.50	8.778	0.000 < 0.05
99th percentile	213.65 ± 32.53	177.81 ± 24.09	6.662	0.000 < 0.05

Values are given as mean ± SD.

ADC, apparent diffusion coefficient.

ROC curve analysis of diagnostic efficacy of full-volume ADC histogram parameters for benign and malignant breast lesions

ROC curves were generated by plotting the eight histogram parameters that were statistically different between the two groups, and the results showed that the AUCs for the differential diagnosis of benign and malignant breast lesions were 0.863 and 0.87, the sensitivities were 88.9% and 82.6%, and specificities were 83.5% and 82%, respectively, when the mean and 50th percentile thresholds were taken as 138.6 and 138.78, respectively (Table 5).

Table 5.

Efficacy of ADC histogram parameters for the diagnosis of benign and malignant breast lesions.

ADC histogram parameters	AUC	95% CI	Threshold	Sensitivity (%)	Specificity (%)
Mean	0.863	0.794∼0.932	138.60	88.91	83.52
Variance	0.697	0.631∼0.818	713.85	79.76	87.11
Skewness	0.498	0.404∼0.593	−0.16	89.04	71.42
10th percentile	0.794	0.710∼0.878	106.55	64.50	92.52
50th percentile	0.870	0.805∼0.936	138.78	82.61	82.03
90th percentile	0.829	0.753∼0.905	179.65	76.71	81.50
99th percentile	0.861	0.676∼0.847	199.73	66.92	80.34

ADC, apparent diffusion coefficient; AUC, area under the curve; CI, confidence interval.

DCE-MRI and DWI-ADC histograms have superior combined diagnostic efficacy for benign and malignant breast lesions

Using the type III outflow curve as a diagnostic criterion for malignant breast lesions, the diagnostic sensitivity was 76.9% and specificity was 80%, with an AUC of 0.823. Using the early intensification rate of stage I > 100% as the diagnostic criterion for malignant breast lesions, the sensitivity was 61.5%, specificity was 80%, and the AUC was 0.723. Using the number of ipsilateral vessels > 3 as a diagnostic criterion for malignant lesions in the breast resulted in a diagnostic sensitivity of 81.6%, specificity of 80.8%, and an AUC of 0.805. The mean and 50th percentile thresholds were taken to be 138.6 and 138.78, respectively, and the AUCs were 0.863 and 0.87, respectively, for the differential diagnosis of benign and malignant breast lesions, with sensitivities of 88.9% and 82.6%, and specificities of 83.5% and 82%, respectively. The combination of TIC type, early enhancement rate in stage I, number and scoring of ipsilateral vessels, mean and 50th percentile quartiles in ADC full-volume histograms yielded a sensitivity of 91.8% and a specificity of 93.3% for discriminating between benign and malignant breast lesions, with an AUC of 0.917, which was higher than the AUC of these four alone (0.823, 0.723, 0.805, 0.863, and 0.87).

Discussion

Breast MRI, a highly sensitive imaging tool for the diagnosis of breast lesions, has gained a series of clinical recognition, including screening of breast lesions, preoperative evaluation for the staging of malignant tumors, and assessment of efficacy (5). DCE-MRI, the mainstay of any breast MRI protocol, has good sensitivity and specificity for breast lesion diagnosis, providing high-resolution morphological signals as well as some functional information on neovascularization. Kinetic analysis of breast lesions can be achieved by the enhancement of the post-time-intensity signal curve (2). According to the expert consensus on breast MR examination and diagnostic rules, the early intensification rate in phase I is defined as the rise of the TIC within 2 min or to the pre-peak curve after the intravenous bolus of contrast. It has been shown that malignant lesions of the breast more often display an outflow type TIC compared to benign lesions (2,16). The present study also showed that type III outflow TICs were more frequently observed in breast malignancies and type I inflow TICs were more frequent in benign tumors.

Fan et al. (17) measured quantitative and semi-quantitative parameters in 21 breast benign lesions and 68 breast malignant lesions, in which the enhancement rate of contrast influx had an AUC of 0.751 for the identification of benign and malignant breast lesions, and its sensitivity was 79.1%, specificity was 70.0%, and accuracy was 77.0%. In this study, there was some variability in the early enhancement rate of the first stage between benign and malignant lesions.

Normal breast tissue and benign breast lesions have a good, but not particularly rich, blood supply, mainly by the axillary artery, intercostal artery, and internal mammary artery. Malignant lesions of the breast mostly have a rich blood supply. There is a close relationship between the feeding artery and the site of the tumor, that is, the site where the tumor is located tends to have a rich blood supply. Both the growth and progression of breast malignancies are dependent on tumor angiogenesis (18,19). A reliable indicator to clinically evaluate tumor angiogenesis is microvessel density (MVD), which can reflect the situation of the balance of angiogenic and antiangiogenic factors within a tumor. However, MVD results can only be obtained by postoperative pathological sections, and MVD is powerless if one wants to preoperatively evaluate the tumor vessels as well as make a judgment on the benign and malignant nature of the mass. Moreover, one study showed that DCE-MRI, as a non-invasive method, can be used to evaluate the microcirculation, such as MVD (20,21). Digital subtraction angiography (DSA) can clearly show the distribution of the mammary feeding arteries; however, DSA was not effective in terms of morphological signs on the breast lesion and the display of the lesion's relationship with the feeding artery. In addition, DSA belongs to invasive examinations, which are difficult to accept in some patients. DCE-MRI vessel MIP 3D reconstruction technique can not only clearly show the blood supply of the breast lesion, but also intuitively show the 3D stereo relationship between the blood supply vessel and the lesion, which is convenient to find the tumor feeding artery and its microcirculation characteristics and has important implications for the diagnosis and differential diagnosis of breast lesions (22 –24).

Numerous studies have shown that ipsilateral blood supply increased in malignant lesions of the breast. The conclusion of Mahfouz et al. (25) for ipsilateral blood supply situation in the diagnosis of malignant lesions of the breast found a sensitivity of 77% and a specificity of 57%, but the study by Carriero et al. (26) confirmed a sensitivity of 72% and specificity of 100%. Thus, we concluded that there were considerable differences in sensitivity and specificity for diagnosing malignant breast lesions based on the ipsilateral blood supply, which may be related to the ratio of the number of benign to malignant lesion groupings in the included cases. The present study concluded that the sensitivity and specificity were 98.5% and 81.3%, respectively, which was somewhat biased from related reports, and the reason for this analysis may be the different composition ratio of benign to malignant lesions in the cases.

Schmitz et al. (27) concluded a sensitivity of 100% and specificity of 87% based on a combined diagnosis of imaging features, TIC as well as vessel count score, and they suggested that a combined analysis of multiple data with imaging methods would greatly improve the diagnostic accuracy. The increased blood supply on one side found in this study, which was pathologically confirmed as malignant lesions, were all of the medium and high blood supply, and the Sardanelli blood supply scores were all > 2 points. Both lesions with no or low blood supply (Sardanelli blood supply score of 0 or 1) were finally pathologically confirmed to be benign, and the number of vessels ipsilateral to both benign and malignant lesions, as well as the Sardanelli blood supply score, were significantly different. However, the results also showed that moderate to high blood supply was not exclusively found in malignant lesions, and eight patients in this study with pathologically confirmed benign lesions were diagnosed as malignant based on the Sardanelli blood supply score; it was possible to determine whether the differential diagnosis between benign and malignant breast lesions can be made solely by ipsilateral blood supply or there was some false diagnosis. The Sardanelli blood supply score, as a more scientific and objective system to evaluate breast blood supply, in practice application may be biased due to the different proportion of specific benign and malignant cases, and in this study, the number of vessels and the Sardanelli blood supply score differed between benign and malignant breast lesions.

DWI is a type of functional imaging of MRI that non-invasively detects the diffusion movement of water molecules within living tissues, and its automatically generated ADC map contains richer texture information. ADC histogram analysis is a method of texture analysis on imaging images, which is analyzed by extracting the pixel value and gray-level distribution of the lesion image, and in turn, obtaining parameters related to the intrinsic properties of the lesion. ADC histogram parameters represent the uniformity and regularity of tissue distribution within a lesion, which can be further assessed by quantifying lesion heterogeneity, such as benign and malignant, pathologic grade, and so on. The whole lesion-based ADC histogram analysis can provide more information than conventional mean ADC values, and in addition to including several percentile ADC values and parameters such as ADC_min, ADC_max, and ADC_mean, parameters reflecting the distribution of ADC values within a lesion, including skewness, kurtosis, and standard deviation, can provide more information reflecting lesion heterogeneity and can be used to better distinguish benign from malignant lesions (28 –30).

In this study, MaZda software was used to perform histogram analysis, and the extractable histogram parameters included mean, variance, skewness, kurtosis, the 1st, 10th, 50th, 90th, and 99th percentiles. The mean value reflects the concentration tendency of the lesion tissues and the average level of the pixel values, and the larger the mean, the brighter the ROI. The variance indicates the mean value and the degree of dispersion of the pixel values, and the larger the pixel value of the lesion, the more the data deviate from the mean value, which can reflect the heterogeneity of the tissue to some extent. The results of the present study showed that the variance in the benign lesion group was larger than that in the malignant lesion group, which indicated that the pixel values of benign lesions were more dispersed than those of malignant lesions. Skewness represents the symmetry of the histogram, with the body of the distribution centered at positive skewness on the right and negative skewness on the left. Our results showed that the histograms of malignant lesions were mainly positively skewed, and the histograms of benign lesions were mainly negatively skewed. The kurtosis reflects the peak of the histogram, and the results of this study showed that there was no significant difference in the comparison of the peak values between the two groups.

The histogram parameters mean and variance, the 1st, 10th, 50th, 90th, and 99th percentiles of the histogram for malignant lesions were lower than those for benign lesions, and the skewness was higher than that for benign lesions, and the mean and 50th percentile AUC for differential diagnosis of benign and malignant breast lesions were 0.863 and 0.87, respectively, with a sensitivity of 88.9% and 82.6%, and a specificity of 83.5% and 82%, respectively. These data suggested that full-volume ADC histogram analysis has high diagnostic efficacy for discriminating between benign and malignant breast lesions. In the present study, the mean and skewness, 10th, 50th, 90th, and 99th percentiles of the histogram parameters were statistically different between the two groups, and the 50th percentile showed the greatest diagnostic efficacy for discriminating between benign and malignant lesions.

In the present study, the four combinations of the type of TIC, the rate of early enhancement in the first phase, the number and scoring of ipsilateral vessels, the mean in the ADC full-volume histogram, and the 50th percentile resulted in a sensitivity of 91.8% and a specificity of 93.3% for the diagnosis of benign and malignant breast lesions, respectively, and improved the diagnostic efficacy of discriminating between benign and malignant breast lesions with an AUC of 0.917, which was higher than the AUC of these four alone (0.823, 0.723, 0.805, 0.863, and 0.87). If these factors could convincingly be used to define a lesion as benign or malignant, patients would not need a second examination or suffer MRI-guided biopsy.

The present study has some limitations. First, the sample size of patients is not large enough. Second, the findings need to be further refined in clinical work. Finally, this study is a single-center study, and more multicenter data are expected in the future.

In conclusion, the type of TIC, the rate of enhancement in the first phase, the number of ipsilateral vessels and the blood supply score, and the ADC full-volume histogram parameters are useful in the differential diagnosis of benign and malignant breast lesions. Combined DCE-MRI and DWI-ADC histograms, morphological features, and other data can significantly improve the differential diagnosis of benign and malignant breast lesions.

Footnotes

Authors’ note

Wen-Jing Li and Guang-Bin Chen are both equal corresponding authors.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) declared the following financial support for the research, authorship, and/or publication of this article: This study was supported by the Department of Education of Hubei Province (No. Q20192101), Health Commission of Hubei Province (No. WJ2019M050), and Shiyan Science and Technology Bureau (No. 2021K66).

ORCID iDs

Wen-Jing Li

Guang-Bin Chen

References

Moy

Elias

Patel

, et al. Is breast MRI helpful in the evaluation of inconclusive mammographic findings? AJR Am J Roentgenol 2009; 193:986–993.

Pinker

Helbich

Morris

EA.

The potential of multiparametric MRI of the breast. Br J Radiol 2017; 90: 20160715.

Allarakha

Gao

Jiang

, et al. Prediction and prognosis of biologically aggressive breast cancers by the combination of DWI/DCE-MRI and immunohistochemical tumor markers. Discov Med 2019; 27:7–15.

Braman

Etesami

Prasanna

, et al. Intratumoral and peritumoral radiomics for the pretreatment prediction of pathological complete response to neoadjuvant chemotherapy based on breast DCE-MRI. Breast Cancer Res 2017; 19:57.

Partridge

Nissan

Rahbar

, et al. Diffusion-weighted breast MRI: Clinical applications and emerging techniques. J Magn Reson Imaging 2017; 45:337–355.

Baltzer

Mann

Iima

, et al. Diffusion-weighted imaging of the breast-a consensus and mission statement from the EUSOBI International Breast Diffusion-Weighted Imaging working group. Eur Radiol 2020; 30: 1436–1450.

Cipolla

Guerrieri

Bonito

, et al. Effects of contrast-enhancement on diffusion weighted imaging and apparent diffusion coefficient measurements in 3-T magnetic resonance imaging of breast lesions. Acta Radiol 2018; 59:902–908.

Ahn

Park

Bang

, et al. Apparent diffusion coefficient histogram in breast cancer brain metastases may predict their biological subtype and progression. Sci Rep 2018; 8:9947.

Horvat

Iyer

Morris

, et al. Histogram analysis and visual heterogeneity of diffusion-weighted imaging with apparent diffusion coefficient mapping in the prediction of molecular subtypes of invasive breast cancers. Contrast Media Mol Imaging 2019; 2019:2972189.

10.

Newitt

Amouzandeh

Partridge

, et al. Repeatability and reproducibility of ADC histogram metrics from the ACRIN 6698 Breast Cancer Therapy Response Trial. Tomography 2020; 6:177–185.

11.

Jin

Zhang

Cheng

, et al. The role of histogram analysis in diffusion-weighted imaging in the differential diagnosis of benign and malignant breast lesions. BMC Med Inform Decis Mak 2020; 20:239.

12.

Liang

Guo

, et al. Diagnostic performance of whole-lesion apparent diffusion coefficient histogram analysis metrics for differentiating benign and malignant breast lesions: a systematic review and diagnostic meta-analysis. Acta Radiol 2020; 61:1165–1175.

13.

Liu

Zong

Wei

, et al. Differentiation between malignant and benign breast masses: combination of semi-quantitative analysis on DCE-MRI and histogram analysis of ADC maps. Clin Radiol 2018; 73:460–466.

14.

Tajima

Kuroki

Tsutsumi

, et al. Pancreatic carcinoma coexisting with chronic pancreatitis versus tumor-forming pancreatitis: diagnostic utility of the time-signal intensity curve from dynamic contrast-enhanced MR imaging. World J Gastroenterol 2007; 13:858–865.

15.

Sardanelli

Iozzelli

Fausto

, et al. Gadobenate dimeglumine-enhanced MR imaging breast vascular maps: association between invasive cancer and ipsilateral increased vascularity. Radiology 2005; 235:791–797.

16.

Partridge

Stone

Strigel

, et al. Breast DCE-MRI: influence of postcontrast timing on automated lesion kinetics assessments and discrimination of benign and malignant lesions. Acad Radiol 2014; 21:1195–1203.

17.

Fan

Chen

Cheng

, et al. Retrospective analysis of the utility of multiparametric MRI for differentiating between benign and malignant breast lesions in women in China. Medicine 2018; 97:e9666.

18.

Weidner

Semple

Welch

, et al. Tumor angiogenesis and metastasis–correlation in invasive breast carcinoma. N Engl J Med 1991; 324:1–8.

19.

Sun

Wang

, et al. Tumor angiogenesis and bone metastasis - Correlation in invasive breast carcinoma. J Immunol Methods 2018; 452:46–52.

20.

Wang

Lin

Peng

LV.

Non-small Cell Lung Cancer: Correlation between Parameters from IVIM, DCE-MRI and Immunohistochemistry Indices MVD, EGFR and Ki67. Chinese Journal of CT and MRI 2017; 15: 60–63.

21.

Jun

Sub

, et al. Correlations of 3T DCE-MRI quantitative parameters with microvessel density in a human-colorectal-cancer xenograft mouse model. Korean J Radiol 2011; 12:722–730.

22.

Jacobs

Wolff

Macura

, et al. Multiparametric and multimodality functional radiological imaging for breast cancer diagnosis and early treatment response assessment. J Natl Cancer Inst Monogr 2015; 2015:40–46.

23.

Gombos

Jagadeesan

Richman

, et al. Magnetic resonance imaging-guided breast interventions: role in biopsy targeting and lumpectomies. Magn Reson Imaging Clin N Am 2015; 23:547–561.

24.

Menezes

Knuttel

Stehouwer

, et al. Magnetic resonance imaging in breast cancer: A literature review and future perspectives. World J Clin Oncol 2014; 5:61–70.

25.

Mahfouz

Sherif

Saad

, et al. Gadolinium-enhanced MR angiography of the breast: is breast cancer associated with ipsilateral higher vascularity? Eur Radiol 2001; 11:965–969.

26.

Carriero

Di Credico

Mansour

, et al. Maximum intensity projection analysis in magnetic resonance of the breast. J Exp Clin Cancer Res 2002; 21:77–81.

27.

Schmitz

Peters

Veldhuis

, et al. Contrast-enhanced 3.0-T breast MRI for characterization of breast lesions: increased specificity by using vascular maps. Eur Radiol 2008; 18:355–364.

28.

Meng

Zhu

, et al. Apparent diffusion coefficient histogram shape analysis for monitoring early response in patients with advanced cervical cancers undergoing concurrent chemo-radiotherapy. Radiat Oncol 2016; 11:141.

29.

Umanodan

Fukukura

Kumagae

, et al. ADC histogram analysis for adrenal tumor histogram analysis of apparent diffusion coefficient in differentiating adrenal adenoma from pheochromocytoma. J Magn Reson Imaging 2017; 45:1195–1203.

30.

Zhang

Chen

Liu

, et al. Assessment of histological differentiation in gastric cancers using whole-volume histogram analysis of apparent diffusion coefficient maps. J Magn Reson Imaging 2017; 45:440–449.