Pär-Ola Bendahl

Background: This study evaluates DL using GEX and preoperatively available clinical data (PreopClinic) to predict SLNM, and explores their potential for guiding axillary surgery and prognostic assessment. Methods: We retrospectively included 6,836 clinically node-negative T1-T2 patients with invasive breast cancer who underwent primary surgery from the SCAN-B. Three DL models—a multilayer perceptron, a pathway-informed sparse neural network, and a transformer—were developed using the development set (n=4,625) and evaluated against XGBoost in the independent test set (n=2,211). Results: The Transformer outperformed other methods for GEX modeling and minimized prior gene selection. In the independent test set, the combined Pre-opClinic+GEX model significantly improved SLNM prediction compared to Pre-opClinic alone (ROC AUC 0.693 vs 0.596, P < 0.001) and identified low-risk patients who might avoid unnecessary SLNB (reduction rate 27.2% at 92.1% sen-sitivity). However, the combined model did not significantly outperform GEX alone. While GEX provided the dominant predictive signal, PreopClinic contributed complementary information with modest numerical gains in clinical utility. Across-subtype training outperformed within-subtype training, particularly in TNBC, where the combined model achieved AUC 0.734 (95% CI: 0.644-0.837). The derived SLNM predictor also provided prognostic information beyond the estab-lished prognostic factors. Although the models were developed primarily using surgical specimen–derived GEX, paired biopsy and surgical-specimen analyses (n=116) demonstrated substantial concordance of transcriptomic patterns and nodal predictions. Conclusion: These findings highlight the Transformer’s robustness against noise and effectiveness in capturing informative transcriptomic features for SLNM pre-diction. The agreement observed between paired biopsy and surgical specimens supports the feasibility of future biopsy-based preoperative applications.

Background Invasive lobular carcinoma (ILC) is a risk factor for positive margins after breast-conserving surgery (BCS). This retrospective observational study examined whether the presence of residual cancer cells in re-excision is associated with ILC or locoregional recurrence. Methods Patients (aged > 18 years) who underwent BCS for invasive or in situ cancer at two sites (2015–2017) were included. Data were retrieved from medical records and the Swedish National Breast Cancer Register. The primary endpoint was residual cancer cells in first re-excision, and secondary endpoints were locoregional recurrence-free interval (LRFi), distant recurrence-free interval, and breast cancer-specific survival. Results In total, 770 patients were included, of whom 132 (17.1%) had positive margins; 130/132 underwent re-excision, and residual cancer cells were detected in 80 (61.5%). Residual cancer cells were weakly and not significantly associated with histological subtype (ILC versus invasive carcinoma of no special type (NST)), but associated with number of positive margins, mammographic lesion size, and pathological extent (all P < 0.05). After a median follow-up time of 7.0 years, locoregional recurrence was diagnosed in 34/721 patients (4.7%). There was no evidence of a difference in LRFi by margin status, residual cancer cells, or histological subtype (all P > 0.4). Molecular subtypes and adjuvant therapies were associated with the risk of locoregional recurrence. Conclusion The presence of residual cancer cells in the first re-excision is not associated with locoregional recurrence risk and only weakly associated with ILC. Number of margins involved and lesion size are predictive of residual cancer cells and might support surgical decision-making at re-excision.

Background: The integration of clinical decision support tools in medical practice is challenging and must be carefully undertaken, especially in cancer management. Noninvasive Lymph Node Status (NILS) is a web-based tool designed to estimate the probability of benign axillary lymph nodes in female patients with breast cancer scheduled for primary surgery. The aim was to identify barriers to NILS adoption in a clinical setting and assess whether intended users can operate the tool without significant errors or difficulties. Additionally, the study aimed to evaluate the appropriateness of result interpretation for decisions to abstain from sentinel lymph node biopsy (SNLB) and measure overall user satisfaction with the tool. Methods: This mixed-methods multicenter on-site qualitative human factor validation study was conducted in a simulated clinical environment, replicating both real-world physical and digital conditions. Based on the identified target user population for the NILS model, twenty physicians comprised the cohort. The study used simulated clinical cases, the System Usability Scale (SUS), and the After-Scenario Questionnaire (ASQ) to evaluate usability and satisfaction. An oral interview was conducted after the evaluations. The study followed a structured protocol, with distinct roles assigned to the test participants, leader, and observer. Results: Twenty physicians from four hospitals, with a median of 9.5 years of specialist practice, participated. Most participants were surgeons (75%, N = 15), while the remaining 25% ( N = 5) were oncologists. Usability scores were high, with a mean SUS score of 89.5 (“excellent”) and a mean ASQ score of 6.3 (Likert scale 1–7). Several interface challenges were identified, including difficulty in locating the reset and information buttons, the ability to enter values outside valid ranges (age and tumor size), and the risk of unintentionally modifying the entered values while scrolling. Conclusions: This study identified the key usability factors, barriers, and facilitators affecting the NILS model implementation. Physicians found the NILS interface easy to use and valued the presented risk estimates of benign axillary lymph node status in their decisions to perform or abstain from SLNB. Study-informed interface redesigns have been implemented. Iterative usability work in clinical settings is crucial for successful implementation. Trial registration: ISRCTN99301435.

Introduction: The complex needs of breast cancer patients should be addressed systematically. This study’s aim was to evaluate a screening-based individualized intervention following breast cancer treatment.Methods: ReScreen is a complex multicenter intervention RCT study in which primary breast cancer patients from southern Sweden were included between 2019 and 2022. Patients were screened using the Distress Thermometer before treatment, with highly distressed patients randomized into the intervention group (n=165) or the control group (n=165). Women with low distress formed an observation group (n=303). The intervention group received individualized rehabilitation that addressed women’s breast cancer treatment-related symptoms and psychosocial needs, and was carried out using a decision support tool. The between-group difference in distress (primary outcome) was evaluated up to 12 months after baseline using linear mixed-model regression analysis, as were differences in physical activity, health-related quality of life, breast cancer symptoms, and resilience.Results: The effect of the intervention on distress was small and not statistically significant. However, effects on physical activity at 6 and 12 months and on fatigue at 12 months were observed. Exploratory analyses revealed some evidence for stronger intervention effects in patients with higher initial distress.Conclusions: Individualized rehabilitation did not change distress but did increase physical activity. Vulnerable patients benefited most from the intervention.Clinical Trial Registration: The ReScreen RCT was registered in the ClinicalTrials.gov registry on 2018/11/15, NCT03434717.

Background: Prediction models for sentinel lymph node (SLN) status could potentially substitute surgical axillary staging in patients with early breast cancer. Several imaging modalities have been used with various feature extraction and selection approaches. This systematic review and meta-analysis aimed to evaluate prediction models for SLN status based on breast imaging in patients with early breast cancer to summarize the current evidence and to identify areas requiring additional research. Methods: The systematic literature search strategy was based on the Population, Intervention, Comparison, and Outcome (PICO) framework: P: female patients with clinically node-negative invasive breast cancer scheduled to undergo primary surgery; I: breast imaging; C: upfront sentinel lymph node biopsy; and O: prediction model performance regarding SLN status. The search was conducted in the PubMed, Embase, Web of Science, Cochrane, and Cumulative Index to Nursing and Allied Health Literature databases in March 2024. The screening of records, data collection, and bias assessments were performed independently by two reviewers. The risk of bias was assessed via the Quality Assessment of Diagnostic Accuracy Studies 2 (QUADAS-2) tool and the Prediction Model Study Risk of Bias Assessment Tool. A meta-analysis was performed using the random-effects model to assess performance and heterogeneity overall and in subgroups. Results: The literature search resulted in the inclusion of 32 articles with 11,464 patients in total. Five imaging categories were included: ultrasound ( n = 8), magnetic resonance imaging (MRI) ( n = 17), mammography ( n = 1), positron emission tomography computed tomography ( n = 1), and multiple modalities ( n = 5). Four studies, assessed as having a high risk of bias, were excluded from the meta-analysis. The meta-analysis revealed heterogeneity in overall performance, except for MRI-based studies, with a pooled area under the curve of 0.85 (95% confidence interval 0.82–0.87). Meta-regression indicated that MRI and model calibration assessment upon validation contributed to heterogeneity. Conclusions: This systematic review and meta-analysis revealed that prediction models using breast imaging—particularly MRI—could serve as a noninvasive alternative to surgical axillary staging in patients with early breast cancer. The results illustrate the heterogeneity between studies and the need for additional high-quality studies. Systematic review registration: PROSPERO CRD42022301852, available at https://www.crd.york.ac.uk/PROSPERO

Background The integration of clinical decision support tools in medical practice is challenging and must be carefully undertaken, especially in cancer management. Noninvasive Lymph Node Status (NILS) is a web-based tool designed to estimate the probability of healthy axillary lymph nodes in female patients with breast cancer scheduled for primary surgery. The aim was to identify barriers to NILS adoption in a clinical setting and assess whether intended users can operate the tool without significant errors or difficulties. Additionally, the study aimed to evaluate the appropriateness of result interpretation for decisions to abstain from sentinel lymph node biopsy (SNLB) and measure overall user satisfaction with the tool. Methods This mixed-methods multicenter on-site qualitative human factor validation study was conducted in a simulated clinical environment, replicating both real-world physical and digital conditions. Based on the identified target user population for the NILS model, twenty physicians comprised the cohort. The study used simulated clinical cases, the System Usability Scale (SUS), and the After-Scenario Questionnaire (ASQ) to evaluate usability and satisfaction. An oral interview was conducted after the evaluations. The study followed a structured protocol, with distinct roles assigned to the test participants, leader, and observer. Results Twenty physicians from four hospitals, with a median of 9.5 years of specialist practice, participated. Most participants were surgeons (75%, N = 15), while the remaining 25% (N = 5) were oncologists. Usability scores were high, with a mean SUS score of 89.5 (“excellent”) and a mean ASQ score of 6.3 (Likert scale 1–7). Several interface challenges were identified, including difficulty in locating the reset and information buttons, the ability to enter values outside valid ranges (age and tumor size), and the risk of unintentionally modifying the entered values while scrolling. Conclusions This study identified the key usability factors, barriers, and facilitators affecting the NILS model implementation. Physicians found the NILS interface easy to use and valued the presented risk estimates of healthy axillary lymph node status in their decisions to perform or abstain from SLNB. Redesign initiatives are ongoing. An iterative process of addressing usability in a clinical setting is crucial for successful implementation. Trial Registration ISRCTN99301435

Background: With a trend toward de-escalation of axillary surgery in breast cancer, prediction models incorporating imaging modalities can help reassess the need for surgical axillary staging. Although mammography is routinely performed for breast cancer imaging, its potential in nodal staging remains underutilized. This study aims to employ advancements in deep learning (DL) to comprehensively evaluate the potential of routine mammograms for predicting lymph node metastasis (LNM) in preoperative clinical settings. Methods: This retrospective study included 1,265 cN0 T1-T2 breast cancer patients, comprising 368 node-positive and 897 node-negative cases, diagnosed from 2009-2017 at three Swedish institutions. Patients diagnosed in 2017 were assigned to the independent test set (n=123, site 2) and the external test set (n=103, site 3), while the remaining patients (n=1,039, site 1 and 2) were used for model development and double cross-validation. A neck module, in conjunction with a ResNet backbone pretrained on unlabeled mammograms, was developed to extract global information from full-breast or region-of-interest (ROI) mammograms by predicting five cancer outcomes. Clinicopathological characteristics were combined with the learned mammogram features to predict LNM collaboratively. The models were evaluated using area under the receiver operating characteristic (ROC) curve (AUC), calibration, and decision curve analysis. Results: Compared to models using only clinical variables, incorporating full-breast mammograms with preoperative clinical variables improved the ROC AUC from 0.690 ± 0.063 (SD) to 0.774 ± 0.057 in the independent test set and from 0.584 ± 0.068 to 0.637 ± 0.063 in the external test set. The combined model showed good calibration and, at sensitivity ≥ 90%, achieved a better net benefit, and a higher sentinel lymph node biopsy reduction rate of 41.7% in the independent test set. Full-breast mammograms showed comparable ability to tumor ROIs in predicting LNM. Conclusion: Our findings underscore that routine mammograms, particularly full-breast images, can enhance preoperative nodal status prediction. They may substitute key postoperative predictors such as pathological tumor size and multifocality, aiding patient stratification before surgery. Interestingly, the added predictive value of mammography was consistent across all sites, whereas the overall performance varied over time periods and sites, likely due to advancements in equipment and procedures.

Purpose: Resilience has been suggested as an important predictor of both physical and mental health-related quality of life in breast cancer patients. However, it is unclear why resilient women handle their diagnosis better, not only mentally, but also physically. The aim of this study was to investigate paths between resilience, physical activity, and mental, physical, and global health-related quality of life in breast cancer patients. Patients and Methods: Structural equation modeling was conducted to evaluate the proposed structural paths using a sample of 638 women with newly diagnosed breast cancer patients from Sweden. Results: Resilience was directly associated with physical activity and mental health-related quality of life. It was indirectly associated with physical functioning, through mental health-related quality of life and physical activity. Resilience was also indirectly associated with global quality of life, through mental health-related quality of life. Conclusion: Mental health support and encouraging physical activity may be especially relevant to enhance all aspects of health-related quality of life early in the breast cancer process. Results should be replicated longitudinally. Plain Language Summary: Receiving a breast cancer diagnosis can be highly disruptive to one’s worldview and sense of normalcy. Many women with breast cancer describe the diagnostic period as even more stressful than receiving treatment. However, some women handle their breast cancer diagnosis better than others. Resilience is a psychological trait that refers to one’s ability to cope with highly stressful events and maintain good wellbeing despite adversity. More resilient women were shown to handle a breast cancer diagnosis better, not only mentally, but also physically. They also describe their quality of life as higher than low resilient women. However, it is unclear why resilient women have not only better mental health, but also physical health. In this study, we addressed this question by examining the relationships between resilience, mental health, physical functioning, physical activity, and patients’ self-reported quality of life. We found that mental health and physical activity explained the relationship between resilience and physical functioning and self-reported quality of life. This suggests that highly resilient women experience less psychological distress following a breast cancer diagnosis and are more inclined to maintain physical activity during this stressful period. Consequently, they may manage to have better physical functioning and evaluate their quality of life as better. This study highlights the importance of psychosocial support and encouraging physical activity in women with breast cancer who may struggle to cope with their diagnosis. This study implies that this can not only alleviate psychological distress, but also aid one’s physical functioning and overall quality of life.

Background and purpose: Tamoxifen remains an important adjuvant treatment in premenopausal patients with hormone receptor-positive breast cancer. Thus, determination of hormone receptors is important. Here, we compare cytosol-based methods, immunohistochemistry (IHC), and gene expression (GEX) analysis for determining hormone receptor status in premenopausal breast cancer patients from a randomised tamoxifen trial, to evaluate their performance in identifying patients that benefit from tamoxifen. Patients and Methods: Premenopausal patients (n=564) were randomised to 2 years of tamoxifen or no systemic treatment. Estrogen receptor (ER) and progesterone receptor (PR) status by protein expression measured by cytosol-based methods and IHC, and mRNA by GEX analysis were compared in 313 patients with available data from all methods. Kaplan Meier estimates and Cox regression were used to evaluate the treatment-predictive value for recurrence-free interval (RFi) and overall survival (OS). Median follow-up for event-free patients was 26 (RFi) and 33 (OS) years. Results: The mRNA data of ESR1 and PGR distributed bimodally, patterns confirmed in an independent cohort. Kappa-values between all methods were 0.76 and 0.79 for ER and PR, respectively. Tamoxifen improved RFi in patients with ER-positive (ER+) or PR-positive (PR+) tumours (Hazard Ratio [HR] and 95% confidence interval [CI]), cytosol-ER+ 0.53 [0.36–0.79]; IHC-ER+ 0.55 [0.38–0.79]; GEX-ER+ 0.54 [0.37–0.77]; cytosol-PR+ 0.49 [0.34–0.72]; IHC-PR+ 0.58 [0.40–0.85]; GEX-PR+ 0.55 [0.38–0.80]). Results were similar for OS. Interpretation: These methods can all identify patients that benefit from 2 years of tamoxifen with equal performance, indicating that GEX data might be used to guide adjuvant tamoxifen therapy.

Background: Surgical sentinel lymph node biopsy (SLNB) is routinely used to reliably stage axillary lymph nodes in early breast cancer (BC). However, SLNB may be associated with postoperative arm morbidities. For most patients with BC undergoing SLNB, the findings are benign, and the procedure is currently questioned. A decision-support tool for the prediction of benign sentinel lymph nodes based on preoperatively available data has been developed using artificial neural network modelling. Methods: This was a retrospective geographical and temporal validation study of the noninvasive lymph node staging (NILS) model, based on preoperatively available data from 586 women consecutively diagnosed with primary BC at two sites. Ten preoperative clinicopathological characteristics from each patient were entered into the web-based calculator, and the probability of benign lymph nodes was predicted. The performance of the NILS model was assessed in terms of discrimination with the area under the receiver operating characteristic curve (AUC) and calibration, that is, comparison of the observed and predicted event rates of benign axillary nodal status (N0) using calibration slope and intercept. The primary endpoint was axillary nodal status (discrimination, benign [N0] vs. metastatic axillary nodal status [N+]) determined by the NILS model compared to nodal status by definitive pathology. Results: The mean age of the women in the cohort was 65 years, and most of them (93%) had luminal cancers. Approximately three-fourths of the patients had no metastases in SLNB (N0 74% and 73%, respectively). The AUC for the predicted probabilities for the whole cohort was 0.6741 (95% confidence interval: 0.6255–0.7227). More than one in four patients (n = 151, 26%) were identified as candidates for SLNB omission when applying the predefined cut-off for lymph node-negative status from the development cohort. The NILS model showed the best calibration in patients with a predicted high probability of healthy axilla. Conclusion: The performance of the NILS model was satisfactory. In approximately every fourth patient, SLNB could potentially be omitted. Considering the shift from postoperatively to preoperatively available predictors in this validation study, we have demonstrated the robustness of the NILS model. The clinical usability of the web interface will be evaluated before its clinical implementation. Trial registration: Registered in the ISRCTN registry with study ID ISRCTN14341750. Date of registration 23/11/2018.