RPS 2002 - Advances in breast intervention

Author Block: T. Hovda¹, K. K. Sahlberg¹, S. Sõber², P. Padrik²; ¹Drammen/NO, ²Tartu/EE
Purpose: This study aimed to explore the potential for polygenetic risk score (PRS) as a tool for risk-based mammographic screening.
Methods or Background: Population-based screening for breast cancer is mainly age-based. This pilot study included 30 women aged 40-50 without prior history of breast cancer, premalignant breast disease or prior genetic testing/counselling. We used a CE-marked clinical test (AnteBC, Antegenes, Estonia) to assess participants’ PRS, and calculate absolute and relative 10-year breast cancer risk. Women were recommended mammographic screening based on the results; standard biennial screening age 50-69 if below average or medium level risk, or earlier start of biennial and/or annual screening if moderate (up to 2x increased) or high (≥2x increased) risk. We also collected information on breast density and family cancer history, and referred the women for further genetic testing for monogenetic pathogenic variants if indicated by family history.
Results or Findings: Mean age was 44.9 years (SD 3.0). Mean absolute 10-year-risk for breast cancer was 1.65% (95% confidence interval [CI] 1.35, 1.91), compared to 1.74% (95% CI 1.63, 1.86) for an average population with the same age distribution. Relative risk ≤1 was observed in 70% (21/30), relative risk 1-2 was observed in 27% (8/30) and relative risk ≥2 was observed in 3% (1/30). We observed no statistically significant difference in relative risk by breast density. Five participants (17%) were recommended further genetic testing due to positive family history.
Conclusion: Polygenetic risk score may be a useful tool for risk-based mammographic screening. Further studies with larger study populations, exploring cost-effectiveness and women's perspective are needed.
Limitations: The small sample size is a limitation, as well as exclusion of women with prior malignant/premalignant breast disease and prior genetic testing, as the results might be biased towards less risk.
Funding for this study: The AnteNOR consortium is sponsoring the study through a grant from the Norwegian Grants Green ICT program.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: The study was approved by the regional ethical committee with approval number: REK 494936.

Author Block: M. M. Nadrljanski, I. B. Krušac, D. Dimitrijevic, L. Raspopovic, A. Djajic, M. Mihajlovic; Belgrade/RS
Purpose: The study aimed to assess contrast-enhanced spectral mammography (CESM) guided breast vacuum-assisted biopsy (VAB), compared to MRI-guided VAB in benign lesion biopsy reduction.
Methods or Background: Consecutive 23 patients with suspicious findings on mammography or digital breast tomosynthesis (n=23; MX-BI-RADS 4), were included prospectively (October 2022 – October 2023), for CESM-guided or MRI-guided VAB. All patients were assessed with CESM. All enhancing lesions were subsequently biopsied with CESM-guided VAB. Non-enhancing lesions were assessed with DCE-MRI (1.5 T/3 T) and the non-mass enhancement (NME) lesions were biopsied with MR-VAB. Non-enhancing lesions on CESM and MRI were downgraded to MRI-BI-RADS 3. Histopathologic correlation was performed for all lesions.
Results or Findings: There were 11 CESM-enhancing lesions (n1=11, 47.83%) and 12 CESM-non-enhancing lesions assessed with DCE-MRI (n2=12, 52.17%). In n1, there were eight malignant lesions (B5, 72.73%: 3 IDC, 1 ILC, 4 DCIS) and three benign lesions (B2, 27.27%). In n2, there were nine NME and three non-enhancing lesions. All nine lesions in n2 were benign (B2, 44.44%: 4 FCC and B3, 55.56%: 2 ADH, 3 FEA). CESM-guided VAB achieved Se=100% (63.1-100.0%); Sp=80% (51.9-95.7%) with NPV 100% and accuracy of 86.9%. MRI-guided VAB in CESM-non-enhancing MRI-NME lesions achieved Se=100% (47.8-100%); Sp=42.9% (9.9-81.6%) and accuracy of 66.7%.
Conclusion: CESM-guided VAB identified all lesions subsequently histopathologically confirmed as malignant. All non-enhancing lesions on CESM (and MRI) were benign. All CESM non-enhancing lesions detectable as NME lesions on MRI, were benign. The absence of enhancement on CESM favours benign lesions. Should larger trials confirm the preliminary findings, CESM-guided VAB may considerably reduce the number of unnecessary biopsies.
Limitations: The limited number of patients (due to the recently introduced method) was an identified limitation.
Funding for this study: No funding was received for this study.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: The study was approved with approval code: 500-01-1335 CESM 03.
Based on the above stated decision, the authorisation was granted for the realisation of the procedure(s).

Author Block: E. Ostillio, P. Costantini, D. Razzini, L. Groenhoff, A. Tambasco, M. Brambilla, A. Gambaro, A. Carriero; Novara/IT
Purpose: To determine whether the Kaiser score can be a valuable tool for radiologists in assessing whether to recommend a biopsy for a BI-RADS 3 lesion.
Methods or Background: Notably, BI-RADS 3 denotes considerable uncertainty in breast nodule diagnoses, situated between benign (BI-RADS 2) and malignant (BI-RADS 4) classifications. The Kaiser score, a magnetic resonance imaging (MRI)-based algorithm developed by Bazar et al., has proven effective for BI-RADS 4 and 5, but its suitability for BI-RADS 3 remains unexplored.
Between January 2016 and August 2023, 100 BI-RADS 3 lesions were studied, excluding 21 patients who did not meet the criteria. These 79 remaining lesions, observed in adult patients with no prior biopsies or treatments, underwent MRI-guided biopsies. Two expert radiologists (with 30 and 25 years of experience) applied the Kaiser score, and the results were compared with histopathological reports.
Results or Findings: Among the biopsied lesions (n=79), 14 were positive, and 65 were negative. Comparing Kaiser score results (Kaiser score: 1-4 = negative; 5-11 = positive) with biopsy findings, the Kaiser score exhibited a sensitivity of 100% (95% CI: 73%-100%) and specificity of 100% (95% CI: 93%-100%).
Conclusion: The strong concordance between the Kaiser score and biopsy outcomes for BI-RADS 3 lesions suggests its potential value in clinical practice for guiding patients towards follow-up or biopsy decisions.
Limitations: This was a single-centre study and, even if it is of a modest number, the population sample could be bigger. Moreover, two expert radiologists analysed the breast lesions through the KS together, after a common agreement, thus impeding us to estimate the KS interobserver agreement.
Funding for this study: Funding was received from the Fondo di Ateneo per la Ricerca - University of Eastern Piedmont.
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information: As it is a retrospective study, no ethics committee approval was required at our institution.

Author Block: E. U. Ekpo¹, I. Hadadi², M. F. F. McEntee³; ¹Sydney/AU, ²Abha/SA, ³Cork/IE
Purpose: The study aimed to optimise the assessment of women recalled at screening and reduce benign biopsy rate.
Methods or Background: We examined the assessment data of 538 women recalled at screening whose mammograms showed features suggestive of breast cancer. Using histopathology results as a reference standard, we examined the diagnostic performance of digital breast tomosynthesis (DBT) and ultrasound as assessment tools. Relative risk analysis and McNemar tests were used to assess and compare the potential for these assessment tools to reduce benign biopsy rates.
Results or Findings: The recall rate was higher in dense breasts. DBT demonstrated higher sensitivity than ultrasound in dense breasts (98.2% versus 80%; p<0.001), but lower specificity (15.4% versus 55%; p<0.001), positive predictive value (PPV) (61.3% versus 71%; p=0.04), and area under the receiver operating characteristic curve (AUC) (0.57 versus 0.67; p=0.001). DBT showed higher sensitivity than ultrasound (99.2% versus 84%; p<0.001) in non-dense breasts, but no differences were observed in specificity, PPV, and AUC. Calcifications were easily detected but overestimated on DBT relative to ultrasound. Ultrasound reduced the benign biopsy rate compared to DBT: BI-RADS A (21% versus 5%; p=0.04); BI-RADS B (23% versus 10%; p=0.003); BI-RADS C (34% versus 7%; p<0.001) and BI-RADS D (39% versus 9%; p<0.001). The number needed to assess to prevent one benign biopsy was significantly lower with ultrasound than DBT in dense breasts: BI-RADS C (1.8 versus 7; p<0.001) and BI-RADS D (1.9 versus 5.1; p=0.03).
Conclusion: In women with dense breasts, DBT has higher sensitivity, but lower specificity and PPV than ultrasound. Compared to DBT, ultrasound reduces the benign biopsy rate for all women.
Limitations: Data were from one facility and 60% of the sample was dense breasts. Also, DBT and ultrasound images were interpreted with knowledge of mammography findings.
Funding for this study: This study was supported by Tour De Cure mid-career funding; University of Sydney, Faculty of Medicine and Health EMCR Emerging Star funding.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: This study was approved by: Sydney Local Health District; Royal Prince Alfred Hospital Human Research Ethics Committee.

Author Block: F. Galati, M. Pasculli, V. Rizzo, G. Moffa, C. Napoletano, B. Cerbelli, C. Catalano, F. Pediconi; Rome/IT
Purpose: The ICE Study is a pilot, prospective, case-control study with the primary aim of characterising the inflammatory response induced by ultrasound-guided tumour cryoablation, in blood samples from breast cancer (BC) patients.
Secondary endpoints are the evaluation of cryoablation efficacy and safety, and imaging prediction of cryoablation effectiveness.
Methods or Background: We enrolled patients with an early-stage BC, scheduled for breast surgery, not eligible for neo-adjuvant therapy and with a cryo-feasible cancer location. Patients enrolled in the cryo-group and in the control-group followed the same therapeutic pathway for the treatment of BC, in terms of blood sampling and surgery. However, the control-group did not undergo cryoablation.
Results or Findings: From July 2022 we enrolled 10 women in the cryo-group and 10 women in the control-group. On the surgical samples of the cryo-group, ultrasound-guided cryoablation caused a steatonecrotic area in all the patients and the treatment was complete in 9 out of 10 patients.
Cryoablation success was evaluated with magnetic resonance imaging (MRI) in 5 patients, with contrast enhanced mammography (CEM) in 4 patients and with breast ultrasound in 1 patient who refused contrast-enhanced imaging. MRI and CEM predicted cryoablation efficacy in 9 out of 9 cases. Regarding circulating markers of cryo-immunological response, the analysis of blood samples revealed a release of HGBM1, which acts as a key mediator of the immune system signalling cellular stress, in both control and cryo-groups. Furthermore, in the cryo-group increased levels of HGBM1 were associated with a significant rise in CD3+ T cells.
Conclusion: Cryoablation is safe, effective and has a role in immune system modulation.
Enhanced imaging (MRI and CEM) can predict procedure success.
Limitations: The limited patient population was an identified limitation.
Funding for this study: "The ICE Study" received funding from the Seed Grant funding programme of the European Society of Radiology (ESR) in collaboration with the European Institute for Biomedical Imaging Research (EIBIR), kindly supported by an unrestricted, non-exclusive grant from GE Healthcare.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: The study obtained the approval of the Institutional Review Board of Sapienza University of Rome (reference: 6528, approved 24.11.2021).

Author Block: S. Carriero, L. Nicosia, A. C. Bozzini, F. Pesapane, A. Latronico, M. Pizzamiglio, E. Cassano; Milan/IT
Purpose: The study aimed to evaluate effectiveness, safety and underestimation of vacuum assisted excision (VAE) compared to surgery in patients with low- and intermediate-grade ductal carcinoma in situ(DCIS).
Methods or Background: From April 2023 to October 2023, all patients with a single cluster of microcalcifications (BI-RADS >3) <15 mm were enrolled. The goal in these patients was to perform VAE of the cluster. The procedure was done using an 8-gauge needle in two stages: first, eight samples were taken with macroscopic excision of the focus, then 4 "cleaning" samples were taken. The samples were categorised into three groups: micro-positive sample (MPS), micro-negative sample (MNS), and cleaning sample (CS). In all cases, the presence of any residual disease within the cleaning area and the rate of post-procedural complications were assessed. A reference clip was always placed at the end of the procedure. All lesions underwent surgery. After surgery, the presence of any residual disease in the surgical specimen was evaluated.
Results or Findings: A total of 98 patients with microcalcifications were prospectively enrolled, resulting in 49 DCI. Seven patients underwent VAE. In 100% of cases (7/7), the underestimation rate was 0%. In 57% (4/7) of cases, no residual disease was found within the cleaning area, and no in-situ pathology was detected in the subsequent surgery. In 43% (3/7), disease was found within the cleaning area, and foci of DCIS were there at surgery. No major post-procedural complications were observed.
Conclusion: Preliminary data have shown that VAE is a safe procedure with no diagnostic underestimation, which could be used in the excision of DCIS as an alternative to surgery, especially in cases where the "cleaning" area is free of DCIS.
Limitations: The major limitation of our study is the limited sample size of the patients
Funding for this study: This research received no external funding.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: The approval for this monocentric prospective study was obtained by the Ethics Committee of IRCCS IEO, Milano.

Author Block: V. Hernström, V. Josefsson, H. Sartor, D. Schmidt, A-M. Larsson, I. Andersson, A. Rosso, O. Hagberg, K. Lang; Malmö/SE
Purpose: To analyse cancer-detection rates and types of detected cancers in the full MASAI-trial study population.
Methods or Background: 105,934 screening participants were randomised (1:1) to AI-supported screening or double-reading without AI (Standard of Care, SoC). In the intervention arm, screening examinations were triaged to low- and high-risk groups based on AI-derived risk scores (ScreenPoint, Transpara v 1.7 score 1–9 and 10, respectively). Low-risk examinations were single read and high-risk examinations double read. Computer-aided detection marks were available for examinations with risk score 8–10. A per protocol analysis on cancer detection in relation to subtype and stage was performed and cancer-detection rates were compared using the Fisher’s exact test with 95% confidence intervals (CI). ClinicalTrials.gov number NCT04838756.
Results or Findings: 53,043 women were allocated to AI-supported screening and 52,872 to SoC. Mean age was 55.1 (SD 10.2) in both arms. 338 cancers were detected with AI-support and 263 with SoC, a cancer-detection rate of 6.4 (95% CI 5.7–7.1) vs. 5.0 (4.4–5.6) per 1000, a ratio of 1.28 (1.09–1.50, p=0.002), an increased detection of 51 invasive and 24 in situ cancers. AI-supported screening detected 204 cancers with non-specific histologic type and 122 cancers with T1c-stage, compared to 155 and 79 with SoC. Notably, 20 more non-luminal A invasive cancers, and 12 more DCIS grade 3 were detected with AI-support.
Conclusion: An AI-supported screen reading protocol resulted in a 28% increase in cancer detection compared to double reading without AI. Subtype and stage of detected cancers suggest that AI can aid in early detection of clinically relevant cancers.
Limitations: It is a single-institution trial.
Funding for this study: Funding was received from: The Swedish Cancer Society; Regional Cancer Centres; Lund University ALF-funds.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: The study was approved by the Swedish Ethical Review Authority (2020-04936).

Author Block: F. Morciano¹, R. Rella¹, M. Conti¹, P. Belli¹, O. Tommasini¹, F. Fornasa², E. Gori¹, G. Romanucci²; ¹Rome/IT, ²Verona/IT
Purpose: The study aimed to assess the diagnostic performance of vacuum-assisted excision (VAE) under stereotactic guidance with "cavity margins shaving" to ensure total lesion removal in patients with ≤1 cm clusters of microcalcifications.
Methods or Background: Patients with ≤1 cm clusters of microcalcifications were eligible. All VAE procedures were performed under stereotactic guidance with a 9G-needle. The first 12 specimens were retrieved into the first sample container (SC); then SC was changed without removing the biopsy needle and cavity margins were shaved with 12 additional specimens into a second SC. Inclusion criteria: complete removal of microcalcifications; B5 or B3 lesion; surgical histopathological examination or imaging follow-up (FUP) ≥24 months. Histopathologic assessment of specimens of the first and second containers was performed separately. Presence of residual lesion in the second SC was compared with surgical excision histology or imaging-FUP evolution, evaluating negative predictive value (NPV) and false-negative rate (FNR).
Results or Findings: A total of 60 lesions were included: 15 B5 and 45 B3. Surgery was performed on 33/60 lesions (all B5 and 18 B3) while the remaining 27/60 B3 lesions were sent to FUP, with one interval change with subsequent surgical excision revealing a ductal carcinoma in situ (DCIS). NPV of the absence of residual lesion in the second container was 81.5% (95% CI: 61.2%-93.0%). FNR was 18.5% (95% CI: 7.0%-38.7%). None of the atypical ductal hyperplasia (ADH) (n=14) or DCIS (n=11) cases were false-negative.
Conclusion: The absence of residual lesion in the "cavity margins shaving" of VAE (using a second SC) can predict complete lesion removal in small clusters of microcalcifications. VAE with "cavity margins shaving" technique is effective in confirming complete lesion removal in small clusters of microcalcifications and seems to be a promising decision-support tool to reduce subsequent surgical excision in selected lesions, such as ADH or DCIS.
Limitations: No limitations were identified.
Funding for this study: No funding was received for this study.
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information: The study is retrospective.

Author Block: G. Irmici¹, C. Depretto¹, L. Rabiolo², G. Della Pepa¹, E. D’Ascoli¹, C. De Berardinis¹, S. Schiaffino³, A. Cozzi³, G. P. Scaperrotta¹; ¹Milan/IT, ²Palermo/IT, ³Lugano/CH
Purpose: B3 breast lesions have uncertain malignant potential, with different risks of upgrade to malignancy at surgery and/or follow-up. This study on B3 lesions presenting as microcalcifications, for which vacuum-assisted biopsy (VAB) represents the standard approach, aimed to investigate the influence of complete or partial lesion removal at VAB on the subsequent upgrade rate (UR).
Methods or Background: For this retrospective monocentric study conducted in a referral centre, we retrieved 165 lesions presenting solely as microcalcifications and being categorised as BI-RADS 4/5 at mammography, then subsequently diagnosed as B3 at VAB (40 ADH, 53 FEA, 40 ALH, 18 PL, and 14 RS) between January 2016 and December 2020. Surgical pathology or at least 3-years follow-up were obtained to determine eventual lesion upgrade to malignancy. The χ², Fisher’s, and Mantel-Haenszel tests were performed to assess if complete lesion removal influenced URs, overall and among different B3 subtypes.
Results or Findings: Complete lesion removal was achieved in 99/165 (60.0%) cases. The rate of complete removal did not significantly differ among B3 subtypes (p=0.092, Bonferroni-adjusted multiple comparisons p≥0.101), ranging from 33.3% of PL (6/18) to 78.6% (11/14) of RS.
The overall UR was 8.5% (95% CI 5.1–13.7%, 14/165), not significantly differing among B3 subtypes (p=0.562).
Conversely, the UR of completely removed lesions (4.0%, 95% CI 1.6–9.9%) was significantly lower than that of partially removed lesions (15.2%, 95% CI 8.4–25.7%, p=0.020).
At stratified analysis according to B3 subtypes, the risk ratio of upgrade among completely and partially removed FEA (0.15, 95% CI 0.01–1.26) was significantly lower (Mantel-Haenszel test p=0.034) than those of ADH (0.36, 95% CI 0.07–1.90) and of ALH (0.74, 95% CI 0.05–10.99).
Conclusion: The UR of B3 lesions is significantly influenced by complete lesion removal, both overall and among different B3 subtypes.
Limitations: The relatively small sample size was an identified limitation.
Funding for this study: No funding was received for this study.
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information: No additional information provided by the submitter.

Author Block: A. Cozzi¹, M. Cao², F. Del Grande¹, N. Sharma³, S. Schiaffino¹; ¹Lugano/CH, ²Milan/IT, ³Leeds/UK
Purpose: The objective of this study was to perform a systematic review and meta-analysis of the upgrade rates to malignancy and the surgical excision rates of B3 lesions treated with vacuum-assisted excision (VAE) after core-needle biopsy (CNB), while also analysing the procedural complication rate of VAE.
Methods or Background: After protocol registration on PROSPERO (CRD42023396663), PubMed and EMBASE were searched for articles published up to 15/04/2023 reporting the use of VAE on B3 lesions diagnosed at CNB, taking surgical pathology and/or follow-up as reference standard. Three readers independently performed article selection and extracted data for the following endpoints: immediate upgrade rate, rate of complications, rates of immediate and long-term surgical excision, upgrade rates after immediate surgical excision and during follow-up. Random-effects meta-analyses of single proportions were performed for each endpoint.
Results or Findings: Sixteen studies (3,941 patients, 3,975 VAE procedures) from six countries, published between 2008 and 2023, were included in quantitative synthesis. The summary immediate upgrade rate at VAE was 4.0% (95% CI 1.4-7.6%). A total of 81 complications were found among 854 patients, the overall summary complication rate being 6.6% (95% CI 1.3-14.6%): 79 (97.5%) were minor complications whereas only two (2.5%) were major complications. Immediate surgical excision was performed in 214 cases, with a 1.5% immediate surgical excision summary rate (95% CI 0.1%-3.6%) and a corresponding 25.4% summary upgrade rate (95% CI 11.0-42.4%). Among 1,375 patients in follow-up, surgical excision was performed in 39 cases, with a 1.4% summary rate of surgical excision during follow-up (95% CI 0.1%-3.4%) and a summary upgrade rate during follow-up of 0.01% (95% CI 0.0-0.6%).
Conclusion: VAE of B3 lesions has low rates of procedural complications, immediate upgrade to malignancy, and subsequent surgical excision, highlighting its promising role as a first-line treatment for high-risk lesions.
Limitations: Substantial-to-high heterogeneity (I²>74.1%) was an identified limitation.
Funding for this study: No funding was received for this study.
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information: This systematic review and meta-analysis did not need any specific ethics committee approval.

Author Block: M. Ş. Aydın, S. Doğan, H. Akgün; Kayseri/TR
Purpose: Breast cancer is a heterogeneous disease. Receptor expression can change throughout the progression of the disease. Clinically significant discordance has been detected between the primary breast tumour (PT) and metastases other than metastatic synchronous axillary lymph nodes (ALN). Therefore, routine biopsy is recommended from metastatic lesion. We aimed to evaluate the discordance between PT and ALN and address the need to concurrently biopsy ALN.
Methods or Background: We prospectively performed core needle biopsies on 95 patients from September 2018 to September 2023 and performed immunohistochemistry (IHC) staining for oestrogen receptor (ER), progesterone receptor (PR), human epidermal growth factor receptor-2 (HER2), and Ki67 in both PTs and ALNs. Cutoff values were >1% for ER/PR and %15 for ki67. HER2 status was determined following ASCO/CAP guidelines. We categorized PTs and ALNs into intrinsic subtypes based on the St Gallen classification. We assessed the consistency using the Cohen's Kappa test.
Results or Findings: Discordance rates between PT and ALN were %6 for ER, %15 for PR, %7 for HER2 and %8 for Ki67. Very good agreement was observed for ER and HER2 (k-values 0.845, 0.812 respectively) and good agreement was observed for PR and Ki67 (k-value 0.666, 0.619). Discordance rates for intrinsic subtypes was found to be %21, and good agreement (k-value 0.724) was observed. There were 10 patients in whom PT was hormone receptor and/or HER2 negative but were positive for these markers in ALN.
Conclusion: There was good statistical consistency between PT and ALN receptor expression. But there were clinically significant discordances that may affect the treatment selection in some patients. Given the lack of complete concordance between PT and ALN, concurrent ALN core needle biopsy may still be considered. Larger studies will be necessary to further define the need of concurrent biopsy.
Limitations: The study included a small sample size.
Funding for this study: No funding was provided for this study.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: This study was approved by the Ethics Committee of Erciyes University Faculty of Medicine under reference number 2018/588.