Accuracy of MRI volume measurements of breast lesions: comparison between automated, semiautomated and manual assessment

Evaluation of Breast Cancer Size Measurement by Computer-Aided Diagnosis (CAD) and a Radiologist on Breast MRI

Purpose: This study aimed to evaluate cancer size measurement by computer-aided diagnosis (CAD) and radiologist on breast magnetic resonance imaging (MRI) relative to histopathology and to determine clinicopathologic and MRI factors that may affect measurements. Methods: Preoperative MRI of 208 breast cancers taken between January 2017 and March 2021 were included. We evaluated correlation between CAD-generated size and pathologic size as well as that between radiologist-measured size and pathologic size. We classified size discrepancies into accurate and inaccurate groups. For both CAD and radiologist, clinicopathologic and imaging factors were compared between accurate and inaccurate groups. Results: The mean sizes as predicted by CAD, radiologist and pathology were 2.66 ± 1.68 cm, 2.54 ± 1.68 cm, and 2.30 ± 1.61 cm, with significant difference (p < 0.001). Correlation coefficients of cancer size measurement by radiologist and CAD in reference to pathology were 0.898 and 0.823. Radiologist’s measurement was more accurate than CAD, with statistical significance (p < 0.001). CAD-generated measurement was significantly more inaccurate for cancers of larger pathologic size (>2 cm), in the presence of an extensive intraductal component (EIC), with positive progesterone receptor (PR), and of non-mass enhancement (p = 0.045, 0.045, 0.03 and 0.002). Radiologist-measured size was significantly more inaccurate for cancers in presence of an in situ component, EIC, positive human epidermal growth factor receptor 2 (HER2), and non-mass enhancement (p = 0.017, 0.008, 0.003 and <0.001). Conclusion: Breast cancer size measurement showed a very strong correlation between CAD and pathology and radiologist and pathology. Radiologist-measured size was significantly more accurate than CAD size. Cancer size measurement by CAD and radiologist can both be inaccurate for cancers with EIC or non-mass enhancement.

Automated breast volume scanner (ABVS) compared to handheld ultrasound (HHUS) and contrast-enhanced magnetic resonance imaging (CE-MRI) in the early assessment of breast cancer during neoadjuvant chemotherapy: an emerging role to monitoring tumor response?

OBJECTIVES

To investigate the role of automated breast volume scanner (ABVS) compared to handheld ultrasound (HHUS) and contrast-enhanced magnetic resonance imaging (CE-MRI) in the early detection of patients with locally advanced breast cancer who are more likely to reach a complete pathological response (pCR) during neoadjuvant chemotherapy (NAC).

METHODS

A single-institution prospective study was performed in patients with histological diagnosis of invasive breast cancer, eligible for NAC, and who were to undergo surgery in our Hospital. Imaging examinations with ABVS, HHUS and CE-MRI were performed at diagnosis (basal time) and after 3 months of chemotherapy (middle time). The tumor size of each lesion was measured at the basal and middle times, and the dimensional variation was reported. Based on this, patients were divided dichotomously by the median value, obtaining "good responders" (goodR) versus "poor responders" (poorR). The results were correlated with the histological assessment (pCR versus No-pCR) with the use of the intergroup comparison of categorical data (Fisher's exact test).

RESULT

A total of 21 patients were included; 5 obtained a pCR (23%). Both the ABVS and the CE-MRI found all 5 patients with pCR in the group of goodR (10 patients), while none of the poorR (11 patients) obtained a pCR [correlation was statistically significant (p 0.01)]. In the HHUS, goodR (10 patients) 1 obtained a pCR while in the poorR (11 patients) 4 obtained a pCR [correlation not statistically significant (p 0.31)].

CONCLUSIONS

ABVS could be a useful tool, appearing to be more reliable than HHUS, and as accurate as CE-MRI, in early detection of patients who could reach a pCR after NAC.

Three-dimensional segmentation of breast masses from digital breast tomosynthesis images

Assessment of three-dimensional (3-D) morphology and volume of breast masses is important for cancer diagnosis, staging, and treatment but cannot be derived from conventional mammography. Digital breast tomosynthesis (DBT) provides data from which 3-D mass segmentation could be obtained. Our method combined Gaussian mixture models based on intensity and a texture measure indicative of in-focus structure, gray-level variance. Thresholding these voxel probabilities, weighted by distance to the estimated mass center, gave the final 3-D segmentation. Evaluation used 40 masses annotated twice by a consultant radiologist on in-focus slices in two diagnostic views. Human intraobserver variability was assessed as the overlap between repeated annotations (median 77% and range 25% to 91%). Comparing the segmented mass outline with probability-weighted ground truth from these annotations, median agreement was 68%, and range was 7% to 88%. Annotated and segmented diameters correlated well with histological mass size (both Spearman's rank correlations [Formula: see text]). The volumetric segmentation demonstrated better agreement with tumor volumes estimated from pathology than volume derived from radiological annotations (95% limits of agreement [Formula: see text] to 11 ml and [Formula: see text] to 41 ml, respectively). We conclude that it is feasible to assess 3-D mass morphology and volume from DBT, and the method has the potential to aid breast cancer management.

Long-term Outcomes of Patients With Giant Intracranial Arteriovenous Malformations

BACKGROUND:

Giant intracranial arteriovenous malformations (AVMs) are rare cerebrovascular lesions that pose management challenges.

OBJECTIVE:

To further clarify outcomes in patients with giant cerebral AVMs managed with conservative or interventional therapies.

METHODS:

We performed a retrospective review of all patients diagnosed with AVMs evaluated at our institution from 1990 to 2013. Patients with a single intracranial AVM &gt;6 cm were included. Patients were divided into 2 groups: conservative management or intervention (microsurgery, radiosurgery, or embolization). Functional outcome was assessed with the modified Rankin Scale (mRS) and compared between the 2 groups.

RESULTS:

A total of 55 patients with giant AVMs were included, and 35 patients (63.6%) had clinical follow-up with a mean of 11.8 years. Spetzler-Martin grades were as follows: grade III, n = 2 (3.6%); grade IV, n = 15 (27.3%); and grade V, n = 38 (69.1%). Twenty-four patients (43.6%) were conservatively managed. The patients in the conservatively managed group had larger AVMs (P &lt; .05) with more frequent involvement of the temporal lobe (P = .02). Five patients (26.3%) in the conservatively managed group and 5 (31.3%) in the intervention group experienced hemorrhage during follow-up, translating to an annualized risk of 2.7% and 4.1%, respectively. No significant difference in risk of first subsequent hemorrhage was observed (P = .78). Despite comparable mRS scores at presentation, we observed a trend toward better outcomes (mRS &lt; 2) in patients undergoing conservative management (P = .06) compared with the intervention group at last follow-up.

CONCLUSION

This study suggests that interventions for giant AVMs should be considered cautiously because hemorrhagic risk is similar regardless of management strategy and functional outcome is likely to be same or better in the conservatively managed population.

Comparison between different imaging techniques in the evaluation of malignant breast lesions: can 3D ultrasound be useful?

PURPOSE

This study was done to assess the feasibility of three-dimensional ultrasonography (3D-US) for volume calculation of solid breast lesions.

MATERIALS AND METHODS

The volumes of 36 malignant lesions were measured using conventional 2D-US, 3D-US and magnetic resonance imaging (MRI) and compared with that obtained with histology (standard of reference). With 2D Ultrasouns, volume was estimated by measuring three diameters and calculating volume with the mathematical formula for spheres. With 3D-US, stored images were retrieved and boundaries of masses were manually outlined; volume calculation was performed with VOCAL software. For MRI, volume measurements were obtained with special software for 3D reconstructions, after each lesion had been manually outlined. Histology measured the three main diameters and the volume was estimated using the mathematical formula for spheres. Interclass correlation coefficient (ICC) and Bland-Altman plots were used to assess agreement between the volumes measured.

RESULTS

ICC indicated that a good level of concordance was identified between 3D-US and histology (0.79). According to the Bland-Altman analysis, limits of agreement of mean differences of the volumes measured with the three imaging modalities were comparable with histology: -2 ÷ 1.5 cm(3) for 3D-US; -2.3 ÷ 1.3 cm(3) for 2D-US and -2.2 ÷ 1.6 cm(3) for MRI.

CONCLUSIONS

3D-US is a reliable method for the volumetric assessment of breast lesions. 3D-US is able to provide valuable information for the preoperative evaluation of lesions.

How to measure breast cancer tumoral size at MR imaging?

OBJECTIVE

To compare the accuracy of different MR sequences to measure tumor size.

METHODS

Eighty-six women (mean age: 53 years (30-78)) who underwent preoperative MRI for breast cancer were included. Maximal diameters of the index tumor (IT) and of the whole extent of the tumor (WET) were measured on T2-weighted (T2W) sequences, on dynamic contrast-enhanced (DCE) T1-weighted (T1W) sequences and on Maximal Intensity Projection (MIP) reconstructions. Agreements with pathological size were evaluated using concordance correlation coefficient (k).

RESULTS

Median pathological size of IT was 20mm (13-25 mm, interquartile range). Median pathological size of the WET was 29 mm (16-50mm, interquartile range). Measurement of IT showed a good concordance with pathological size, with best results using T2W (k = 0.690) compared to MIP (k = 0.667), early-subtracted DCE frame (k = 0.630) and early-native DCE frame (k = 0.588). IT was visible on T2W in 83.7% and accurately measured within 5mm in 69.9%. Measurement of WET was superior using early-subtracted DCE frame (k = 0.642) compared to late-native frame (k = 0.635), early-native frame (k = 0.631), late-subtracted frame (k = 0.620) and MIP (k = 0.565). However, even using early-subtracted frame, WET was accurately measured within 5mm only 39.3%.

CONCLUSION

If visible, IT size is best measured on T2W with a good accuracy (69%) whereas WET is best estimated on early-subtracted DCE frame. However, when adjacent additional sites exist around IT, suspected surrounding disease components need to be proved by pathological analysis.

Diffusion-weighted magnetic resonance imaging for pretreatment prediction and monitoring of treatment response of patients with locally advanced breast cancer undergoing neoadjuvant chemotherapy

BACKGROUND

For patients with locally advanced breast cancer (LABC) undergoing neoadjuvant chemotherapy (NACT), the European Guidelines for Breast Imaging recommends magnetic resonance imaging (MRI) to be performed before start of NACT, when half of the NACT has been administered and prior to surgery. This is the first study addressing the value of flow-insensitive apparent diffusion coefficients (ADCs) obtained from diffusion-weighted (DW) MRI at the recommended time points for pretreatment prediction and monitoring of treatment response.

MATERIALS AND METHODS

Twenty-five LABC patients were included in this prospective study. DW MRI was performed using single-shot spin-echo echo-planar imaging with b-values of 100, 250 and 800 s/mm(2) prior to NACT, after four cycles of NACT and at the conclusion of therapy using a 1.5 T MR scanner. ADC in the breast tumor was calculated from each assessment. The strength of correlation between pretreatment ADC, ADC changes and tumor volume changes were examined using Spearman's rho correlation test.

RESULTS

Mean pretreatment ADC was 1.11 + or - 0.21 x 10(-3) mm(2)/s. After 4 cycles of NACT, ADC was significantly increased (1.39 + or - 0.36 x 10(-3) mm(2)/s; p=0.018). There was no correlation between individual pretreatment breast tumor ADC and MR response measured after four cycles of NACT (p=0.816) or prior to surgery (p=0.620).

CONCLUSION

Pretreatment tumor ADC does not predict treatment response for patients with LABC undergoing NACT. Furthermore, ADC increase observed mid-way in the course of NACT does not correlate with tumor volume changes.