Multi-center evaluation of artificial intelligent imaging and clinical models for predicting neoadjuvant chemotherapy response in breast cancer.

BACKGROUND: Neoadjuvant chemotherapy (NAC) plays an important role in the management of locally advanced breast cancer. It allows for downstaging of tumors, potentially allowing for breast conservation. NAC also allows for in-vivo testing of the tumors' response to chemotherapy and provides important prognostic information. There are currently no clearly defined clinical models that incorporate imaging with clinical data to predict response to NAC. Thus, the aim of this work is to develop a predictive AI model based on routine CT imaging and clinical parameters to predict response to NAC. METHODS: The CT scans of 324 patients with NAC from multiple centers in Singapore were used in this study. Four different radiomics models were built for predicting pathological complete response (pCR): first two were based on textural features extracted from peri-tumoral and tumoral regions, the third model based on novel space-resolved radiomics which extract feature maps using voxel-based radiomics and the fourth model based on deep learning (DL). Clinical parameters were included to build a final prognostic model. RESULTS: The best performing models were based on space-resolved and DL approaches. Space-resolved radiomics improves the clinical AUCs of pCR prediction from 0.743 (0.650 to 0.831) to 0.775 (0.685 to 0.860) and our DL model improved it from 0.743 (0.650 to 0.831) to 0.772 (0.685 to 0.853). The tumoral radiomics model performs the worst with no improvement of the AUC from the clinical model. The peri-tumoral combined model gives moderate performance with an AUC of 0.765 (0.671 to 0.855). CONCLUSIONS: Radiomics features extracted from diagnostic CT augment the predictive ability of pCR when combined with clinical features. The novel space-resolved radiomics and DL radiomics approaches outperformed conventional radiomics techniques.

Overlap of high-risk individuals predicted by family history, and genetic and non-genetic breast cancer risk prediction models: implications for risk stratification.

BACKGROUND: Family history, and genetic and non-genetic risk factors can stratify women according to their individual risk of developing breast cancer. The extent of overlap between these risk predictors is not clear. METHODS: In this case-only analysis involving 7600 Asian breast cancer patients diagnosed between age 30 and 75 years, we examined identification of high-risk patients based on positive family history, the Gail model 5-year absolute risk [5yAR] above 1.3%, breast cancer predisposition genes (protein-truncating variants [PTV] in ATM, BRCA1, BRCA2, CHEK2, PALB2, BARD1, RAD51C, RAD51D, or TP53), and polygenic risk score (PRS) 5yAR above 1.3%. RESULTS: Correlation between 5yAR (at age of diagnosis) predicted by PRS and the Gail model was low (r=0.27). Fifty-three percent of breast cancer patients (n=4041) were considered high risk by one or more classification criteria. Positive family history, PTV carriership, PRS, or the Gail model identified 1247 (16%), 385 (5%), 2774 (36%), and 1592 (21%) patients who were considered at high risk, respectively. In a subset of 3227 women aged below 50 years, the four models studied identified 470 (15%), 213 (7%), 769 (24%), and 325 (10%) unique patients who were considered at high risk, respectively. For younger women, PRS and PTVs together identified 745 (59% of 1276) high-risk individuals who were not identified by the Gail model or family history. CONCLUSIONS: Family history and genetic and non-genetic risk stratification tools have the potential to complement one another to identify women at high risk.

Breast-related extranodal Rosai-Dorfman disease presenting as subcutaneous masses with thick hyperechoic rim, with review of the literature.

Rosai-Dorfman disease (RDD) is a rare, idiopathic histiocytic proliferative disorder. We report two cases of RDD related to the breast which showed common distinctive imaging characteristics which can help facilitate accurate diagnosis and appropriate management.

Would Removal of All Ultrasound Abnormal Metastatic Lymph Nodes Without Sentinel Lymph Node Biopsy Be Accurate in Patients with Breast Cancer with Neoadjuvant Chemotherapy?

LESSONS LEARNED: Removal of sonographically abnormal (up to 3) metastatic clipped nodes, without sentinel lymph node biopsy, could accurately predict axillary status in breast cancer patients receiving neoadjuvant chemotherapy. ypT and the first clipped node status were statistically significant factors for nodal pathologic complete response. This novel approach requires validation in larger studies. BACKGROUND: In patients who have node-positive breast cancer, neoadjuvant chemotherapy could result in nodal pathologic complete response (pCR) and avoid an axillary lymph node dissection (ALND). Axillary staging, in such cases, can be performed using targeted axillary dissection (TAD) with a low false negative rate. However, identification of sentinel lymph nodes (SLNs) after chemotherapy can be difficult, and currently, it is the standard to remove only one clipped node in TAD. We aimed to determine if removal of all sonographically abnormal metastatic clipped nodes, without SLN biopsy, could accurately predict the axillary status post neoadjuvant chemotherapy. METHODS: Patients with breast cancer with one to three sonographically abnormal metastatic axillary nodes were prospectively recruited. Each abnormal node had histology and clip insertion before neoadjuvant chemotherapy. After chemotherapy, the patients underwent removal of clipped nodes using the Skin Mark clipped Axillary nodes Removal Technique (SMART) and ALND. RESULTS: Fourteen patients were recruited, having a total of 21 sonographically abnormal metastatic nodes, with nine, three, and two patients having 1, 2, and 3 malignant nodes clipped, respectively. Mean age was 55.5 years; 92.9% and 57.1% of patients had invasive ductal carcinoma and grade III tumors, respectively; and 35.7% patients achieved nodal pCR. The first clipped node predicted the axillary status with a false negative rate of 7.1%. Adding to this another second clipped node, the false negative rate was 0%. Pathologic tumor staging after neoadjuvant chemotherapy (ypT) (p = .0390) and the first clipped node pathological response status (p = .0030) were statistically significant predictors for nodal pCR. CONCLUSION: Removal of sonographically abnormal metastatic clipped nodes using SMART, without sentinel lymph node biopsy, could accurately predict axillary status. This finding needs validation in larger studies.

Evaluating the preoperative breast cancer characteristics affecting the accuracy of axillary ultrasound staging.

We evaluate the preoperative breast cancer (BC) characteristics that affect the diagnostic accuracy of axillary ultrasound (US) and determine the reliability of US in the different subgroups of BC patients. Axillary US assessments in women with invasive BC diagnosed between 2009 and 2016 in a single institution were retrospectively reviewed. The diagnostic accuracy of axillary US was obtained using surgical nodal histology as the gold standard. Preoperative breast tumor sonographic and histological factors affecting axillary US diagnostic accuracy were examined. Of the 605 newly diagnosed invasive BC cases reviewed, 251 (41.5%) had nodal metastases. Axillary US sensitivity was 75.7%, specificity 92.9%, positive predictive value 88.4%, negative predictive value 84.4%, and false-negative rate 24.3%. Lower US sensitivity was seen with invasive lobular cancer (ILC) (P = .043), grade I/II, (P = .021), unifocal (P = .039), and smaller tumors (P < .001). US specificity was lower in grade III (P < .001), estrogen receptor (ER)-negative (P < .001), progesterone receptor (PR)-negative (P = .004), HER2-positive (P = .015), triple-negative (P = .001), and larger breast tumors (P < .001). US has moderate sensitivity and good specificity in detecting metastatic axillary lymph nodes. Based on preoperative cancer characteristics, US was less sensitive for nodal metastases from ILC, unifocal, lower grade, and smaller breast tumors. It was also less specific in grade III, ER-negative, PR-negative, HER2-positive, triple-negative, and larger breast tumors. Caution is suggested in interpreting the US axillary findings of patients with these preoperative tumor features.

Comparative Study of Pattern-Based Versus Size Ratio Ultrasound Strain Elastographic Techniques on Breast Masses.

OBJECTIVES: The aim of this study was to compare the diagnostic performance of strain elastographic techniques involving pattern-based elastography versus size ratio elastography on breast masses. METHODS: Female patients presenting to the radiology department for ultrasound (US)-guided biopsies, fine-needle aspirations, or localizations of breast masses were invited to undergo US elastographic assessments of the lesions before the breast procedures. The masses were evaluated for color score and color map size ratio elastographic assessments with one US machine and size ratio and grayscale strain pattern elastographic assessments with another machine. With the histopathologic diagnosis as the reference standard, the sensitivity, specificity, and area under the receiver operating characteristic curve were compared among the 4 groups. RESULTS: A total of 154 breast masses from 139 women were assessed. The sensitivity and specificity were 0.615 (95% confidence interval, 0.470-0.747) and 0.824 (0.736-0.892), respectively for the color score, 0.962 (0.868-0.995) and 0.765 (0.670-0.843) for the size ratio, 0.904 (0.790-0.968) and 0.745 (0.649-0.826) for the color map size ratio, and 0.635 (0.490-0.764) and 0.422 (0.324-0.523) for the grayscale strain pattern. The areas under the receiver operating characteristic curve were 0.729 for the color score, 0.874 for the size ratio (P = .001 with the color score as a reference), 0.836 for the color map size ratio (P = .002), and 0.556 for the grayscale strain pattern (P < .001). CONCLUSIONS: Size ratio elastography had higher sensitivity and better diagnostic accuracy compared to color score elastography. This superiority was demonstrated on two different US systems, and size ratio elastography should thus be considered over color score elastography in the adjunctive US assessment of breast lesions.

MRI kinetics with volumetric analysis in correlation with hormonal receptor subtypes and histologic grade of invasive breast cancers.

OBJECTIVE. The aim of this study was to assess whether computer-assisted detection-processed MRI kinetics data can provide further information on the biologic aggressiveness of breast tumors. MATERIALS AND METHODS. We identified 194 newly diagnosed invasive breast cancers presenting as masses on contrast-enhanced MRI by a HIPAA-compliant pathology database search. Computer-assisted detection-derived data for the mean and median peak signal intensity percentage increase, most suspicious kinetic curve patterns, and volumetric analysis of the different kinetic patterns by mean percentage tumor volume were compared against the different hormonal receptor (estrogen-receptor [ER], progesterone-receptor [PR], ERRB2 (HER2/neu), and triple-receptor expressivity) and histologic grade subgroups, which were used as indicators of tumor aggressiveness. RESULTS. The means and medians of the peak signal intensity percentage increase were higher in ER-negative, PR-negative, and triple-negative (all p ≤ 0.001), and grade 3 tumors (p = 0.011). Volumetric analysis showed higher mean percentage volume of rapid initial enhancement in biologically more aggressive ER-negative, PR-negative, and triple-negative tumors compared with ER-positive (64% vs 53.6%, p = 0.013), PR-positive (65.4% vs 52.5%, p = 0.001), and nontriple-negative tumors (65.3% vs 54.6%, p = 0.028), respectively. A higher mean percentage volume of rapid washout component was seen in ERRB2-positive tumors compared with ERRB2-negative tumors (27.5% vs 17.9%, p = 0.020). CONCLUSION. Peak signal intensity percentage increase and volume analysis of the different kinetic patterns of breast tumors showed correlation with hormonal receptor and histologic grade indicators of cancer aggressiveness. Computer-assisted detection-derived MRI kinetics data have the potential to further characterize the aggressiveness of an invasive cancer.

Use of axillary ultrasound to guide breast cancer management in the genomic assay era.

INTRODUCTION: Chemotherapy is conventionally offered to non-stage IV breast cancer patients with metastatic nodes. However, the RxPONDER trial showed that chemotherapy can be omitted in selected patients with 1-3 metastatic nodes if the 21-gene assay recurrence score is ≤25. We aimed to investigate if axillary ultrasound can identify this group of patients with limited nodal burden so that they can undergo upfront surgery followed by gene assay testing, to potentially avoid chemotherapy. METHODS: T1-3, node positive, hormone receptor-positive and HER2-negative breast cancer patients ≥50 years old with axillary lymph node dissection (ALND) were reviewed from 2 centres. Patients with neoadjuvant chemotherapy and bilateral cancers were excluded. Number of ultrasound-detected abnormal axillary nodes, demographic and histological parameters were correlated with the number of metastatic nodes found on ALND. RESULTS: 138 patients were included, 59 (42.8%) and 79 (57.2%) patients had 1-3 and >3 metastatic nodes on ALND respectively. On logistic regression and ROC analysis, the number of ultrasound-detected abnormal nodes was significant (p < 0.001) for predicting limited nodal burden (ROC AUC = 0.7135). Probabilities of <4 metastatic nodes with ultrasound cut-offs of 5, 6 and 8 abnormal nodes were 0.057, 0.026 and 0.005 respectively, with 100% specificity. CONCLUSION: A cut-off of ≤5 ultrasound-detected abnormal nodes can distinguish between patients with limited versus high nodal burden, with high specificity. Hence, incorporating the number of abnormal ultrasound-detected nodes into clinical practice may prove useful in guiding between upfront surgery and gene assay testing or neoadjuvant chemotherapy in this group of patients.

Oncologic outcomes in breast cancer patients with metastatic nodes and pathological nodal response following neoadjuvant chemotherapy without axillary dissection: a literature review.

Background and Objective: With promising nodal pathological complete response (pCR) after neoadjuvant chemotherapy, the role of axillary lymph node dissection (ALND) was questioned. While there is much data on the accuracy of axillary staging after neoadjuvant chemotherapy to predict nodal pCR, limited data on the oncological safety of omitting ALND exists. We aim to review the oncological safety of omitting ALND in patients with initially metastatic nodes achieving nodal pCR, based on axillary staging, following neoadjuvant chemotherapy. Methods: A PubMed search of relevant articles from 1st January 2013 to 15th September 2022 was performed. Studies with duplication of patients, ALND only, no oncological details, initially N0 patients only and patients without nodal pCR were excluded. Key Content and Findings: Fifteen studies with 1,515 eligible patients (range, 29-242 patients) were analysed. There was heterogeneity of patients among the included studies and patients had varying TN staging, making selection criteria for omission of ALND inconclusive. Sentinel lymph node biopsy (SLNB) was the most studied mode of axillary staging in 1,416 (93.5%) patients, though 35.7% had <3 SLN harvested. On average median follow-up of 52.8 months (range, 9-110 months), axillary recurrence ranged from 0% to 3.4%. There was limited data on survival outcomes. Conclusions: In node positive breast cancer patients who achieved nodal pCR after neoadjuvant chemotherapy, the axillary recurrence rate was low without ALND. However, survival data was limited. The selection criteria and ideal axillary staging technique for patients, who are suitable for axillary preservation, are unclear. More prospective studies with longer follow-up, providing survival data, are needed.

Extensively infarcted breast cancer.

This is a case of a tumour that appeared largely unviable after near complete infarction. The lesion presented as a regular shaped mass with cystic appearance lacking definitive malignant radiological signs. Together with the initial non-diagnostic histological result, this could have easily led to a missed diagnosis of cancer.

Encapsulated papillary carcinoma of the breast: An institutional case series and literature review.

BACKGROUND: Encapsulated papillary carcinoma of the breast is rare, making difficult diagnosis and resulting in patients undergoing excision biopsy before definitive surgery. Evidence-based guidelines are sparse. We would like to further elucidate the clinicopathological, treatment and survival outcomes. MATERIALS AND METHODS: 54 patients identified, with a median follow up duration of 48 months. Patients' demographics, radiological and clinicopathological characteristics, treatment, adjuvant therapies as well as survival data were analysed. RESULTS: 18 (33.3%) cases were pure EPC, 12 (22.2%) were EPC associated with ductal carcinoma in situ (DCIS) and 24 (44.4%) cases had concurrent invasive ductal carcinoma. EPCs were more likely to present as a solid-cystic mass on sonography (63.8%), regular-shaped (oval or round) (97.9%), lack spiculations (95.7%) and lack suspicious microcalcifications (95.6%). Median tumour size was largest in the EPC with IDC group (18.5 mm). 2 patients developed loco-regional recurrence. Overall survival is good for EPCs of all subtypes. CONCLUSION: EPC is a rare tumour with excellent prognosis.

Determining the benefit of neoadjuvant chemotherapy in reduction of axillary dissection rates in Z0011 trial cohort with high nodal burden.

Background: In breast cancer patients fulfilling the Z0011 trial criteria, axillary lymph node dissection (ALND) is reserved for patients with a high nodal burden of ≥3 metastatic nodes. In this group of patients, to avoid an ALND, neoadjuvant chemotherapy (NACT) could be given instead to achieve nodal pathological complete response (pCR). However, the benefit of NACT in achieving nodal pCR and avoiding ALND in this group of patients is unknown. We aimed to determine the nodal pCR rate in this group of patients who otherwise would have needed an ALND. Methods: cT1-2N0 breast cancer patients, with histologically proven nodal metastasis, who underwent NACT were identified from a prospectively maintained database. The sonographic criteria of ≥3 abnormal nodes, which has been reported as highly predictive of high nodal burden, was then used to identify the high nodal burden group. Nodal pCR was determined based on the ALND following NACT. Results: Twenty-four patients with high nodal burden were identified. Mean age was 55.2 years. 91.7% had invasive ductal carcinoma and 29.2% had grade III cancer. 54.2% achieved nodal pCR which was associated with ypT (P=0.006). Nodal pCR was 75%, 70% and 30% in the triple negative, human epidermal growth factor receptor2 (HER2) positive and ER/PR+HER2- tumors, respectively. Conclusions: In the postulated T1-2 breast cancer patients with high nodal burden, needing an upfront ALND, NACT could result in nodal pCR of 54.2%, with higher pCR in certain subtypes. Hence, to minimize ALND risk, NACT should be offered in this high nodal burden group.

Shear wave versus strain elastography of breast lesions-The value of incorporating boundary tissue assessment.

PURPOSE: This prospective study compares the diagnostic performance of strain elastography (SE) based on the elastogram to B-mode (E/B) size ratio to shear-wave elastography (SWE) incorporating assessment of boundary tissue stiffness. METHODS: In this prospective study, 81 women consented to undergo ultrasound (US) elastographic assessment prior to biopsy of their mass(es). A total of 101 masses were evaluated with SE and SWE using a single machine (Siemens Acuson S2000). The histopathologic diagnosis was used as the reference standard. The sensitivity, specificity and area under curve (AUC) were obtained for SE and SWE. RESULTS: The sensitivity and specificity of SE using the E/B ratio was 94.7% (95% CI: 87.6 100) and 81.0% (95% CI: 71.3 90.1) respectively. For SWE the sensitivity and specificity was 73.7% (95% CI: 59.7 87.7) and 82.5% (95% CI: 73.2 91.9) respectively. The AUC for SE was higher (0.878) compared to SWE (0.697). CONCLUSION: SE elastography using the E/B ratio yields diagnostic performance which is superior to SWE incorporating assessment of peripheral stiffness.