Annals of Surgical Oncology最新文献

Background: Intraoperative diagnosis of visceral pleural invasion (VPI) during video-assisted thoracoscopic surgery (VATS) remains challenging. This study aimed to develop and validate a deep learning-based model to improve diagnostic accuracy and guide surgical decision-making.

Methods: Thoracoscopic videos and clinical data from 346 patients (3367 images, 2015-2024) in one hospital were divided into training, validation, and internal-test sets (7:2:1), whereas data from 53 patients (1274 images) in two other hospitals formed the external-test set. A spatial dropout-based Residual Convolutional Neural Network (VPI-Net) was developed for estimating patients' VPI status and VPI risk score (VPIscore). The model's performance was compared against intraoperative estimations by surgeons and preoperative assessments by radiologists.

Results: The VPI-Net model demonstrated significantly higher area under the curve (AUC, 0.84-0.94) and accuracy (79.67-88.68%,) than two surgeons and one radiologist across all cohorts (p < 0.05). Additionally, the VPI-Net model outperformed human experts in sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) across all cohorts. A lower VPIscore (VPIscoreL) was significantly correlated with longer overall survival (OS), relapse-free survival (RFS), and time to progression (TTP) than a higher VPIscore (VPIscoreH) (all p < 0.001). Similar results were observed in patients who had small tumors, with those who had VPIscoreH exhibiting significantly worse RFS and TTP than those with VPIscoreL (RFS [p = 0.012], TTP [p = 0.035]). The VPIscoreL patients had a significantly longer TTP (p = 0.03) than the VPIscoreH patients after sublobectomy.

Conclusion: The proposed model enables satisfactory intraoperative identification of VPI, potentially improving patient outcomes during VATS.

Background: Prostate cancer remains a leading cause of cancer mortality, yet current risk stratification systems inadequately integrate multidimensional tumor characteristics. This study aims to develop a deep learning-based multimodal prognostic model that combines genomic signatures, histomorphological features from whole-slide imaging (WSI), and clinical parameters to improve risk stratification and therapeutic decision-making.

Materials and methods: We constructed a deep learning framework using two independent cohorts: The Cancer Genome Atlas (TCGA) for training and the Prostate, Lung, Colorectal and Ovarian (PLCO) trial for external validation. The model sequentially extracted 768-dimensional histopathological features from hematoxylin and eosin (H&E)-stained WSIs, predicted genomic scores, and histopathological scores, and integrated these with clinical variables via Cox regression to generate a multimodal prognostic score. Performance was evaluated through Kaplan-Meier analysis, Harrell's concordance index, and multivariate Cox regression.

Results: The prognostic score demonstrated superior prognostic accuracy (C-index: 0.774) compared with unimodal scores (genomic, histopathological, clinical) and NCCN risk stratification (C-index: 0.706-0.746, p < 0.05). High-risk patients identified by the multimodal model had significantly shorter progression-free intervals (HR = 9.088, 95% CI 6.033-13.691, p < 0.0001) and prostate cancer-specific mortality (HR = 8.787, 95% CI 3.238-23.847, p < 0.0001). Subgroup analysis within NCCN high-risk patients revealed distinct survival trajectories (p < 0.001). Gene set enrichment linked multimodal scores to tumor-relevant pathways.

Conclusions: This integrative model eliminates reliance on genomic testing by computationally inferring genomic features from histopathology, while significantly enhancing prognostic precision. By stratifying heterogeneous patient populations and refining existing NCCN classifications, the prognostic score offers clinical potential for personalized risk assessment and optimized therapeutic strategies.

Background: The oncologic safety of minimally invasive total gastrectomy (MITG) compared with open total gastrectomy (OTG) for locally advanced gastric cancer remains unclear. This study aimed to evaluate the long-term oncologic outcomes of MITG compared with OTG for locally advanced gastric cancer.

Methods: From 2007 to 2019, 1319 OTG and 348 MITG patients with locally advanced gastric cancer were retrospectively analyzed. The long-term oncologic outcomes of MITG and OTG were compared using propensity score-matching (PSM).

Results: After PSM, clinicopathologic features were well-balanced. The MITG procedure showed less blood loss but a longer operative time. The rates of complications classified as Clavien-Dindo grade ≥III were comparable in the two groups (OTG 12.4% vs. MITG 10.6%; P = 0.470), including anastomotic leakage (OTG 2.9% vs. MITG 4.7%; P = 0.230). The 5 year overall survival rate was 83.0% in the OTG group (95% confidence interval [CI], 78.4-86.7%) and 87.3% in the MITG group (95% CI, 83.2-90.5%) (P = 0.398). The hazard ratio (HR) for death in the MITG group compared with the OTG group was 0.88 (95% CI, 0.61-1.27; P = 0.505). The 5-year relapse-free survival was 71.0% in the OTG group (95% CI, 64.8-76.4%) and 72.5% in the MITG group (95% CI, 66.6-77.5%) (P = 0.895). The HR for recurrence in the MITG group compared with the OTG group was 1.08 (95% CI, 0.80-1.46; P = 0.633).

Conclusion: For locally advanced gastric cancer, MITG demonstrated long-term oncologic outcomes similar to those of OTG. Therefore, MITG could be an oncologically safe option for locally advanced gastric cancer, although randomized studies are needed to confirm this finding.

Cancer surgery is an integral component of the cancer care delivery pathway. Surgical intervention is required across the entire spectrum of cancer care, ranging from screening to treatment and palliation. More than 80% of patients with solid tumors will require surgical intervention at least once, with some needing it multiple times. However, access to safe, timely, and affordable cancer surgery remains beyond the reach of many, especially in low- and middle-income countries (LMICs). In the future, LMICs will account for a rising proportion of the global cancer burden, with a proportionate rise in demand for cancer surgical services. Persistent workforce shortages, fragmented training pathways, limited investment in global cancer surgery research, and financial toxicity further exacerbate these disparities. Addressing these gaps requires comprehensive, context-specific, geographically relatable, and resource-stratified cancer care strategies. This review aims to highlight some of the key literature for improving access to cancer surgery globally, with an emphasis on LMICs.

Background: Cachexia is associated with worse postoperative outcomes, but the added role of neoadjuvant therapy (NAT) is unclear. This study evaluated whether preoperative cachexia and NAT act as a "double jeopardy" after pancreatoduodenectomy.

Patients and methods: A nationwide observational cohort study was conducted using the Norwegian NORGAST registry (2016-2023). Adults undergoing pancreatoduodenectomy for malignancy were included. Cachexia was defined by consensus weight-loss criteria. Modified Poisson and Cox models (with a cachexia and NAT interaction term) estimated adjusted risk ratios (aRR) for textbook outcome (TO), prolonged length-of-stay (LOS), and adjusted hazard ratios (aHR) for overall survival.

Results: Of 1424 patients undergoing pancreatoduodenectomy, cachexia was present in 588 (41.3%). Having cachexia was associated with higher TO (aRR 1.28, 95% CI 1.13-1.46) with effect modification by body mass index (BMI) (interaction P = 0.047). Patients with cachexia had a lower risk of prolonged LOS (aRR 0.64, 95% CI 0.51-0.80). Cachexia was not independently associated with overall survival (aHR 1.15, 95% CI 0.97-1.36). NAT was associated with a higher hazard of death (aHR 1.44, 95% CI 1.09-1.92), likely reflecting confounding by indication. No statistically significant interaction between cachexia and NAT was observed for TO (P = 0.277) or for survival (P = 0.863).

Conclusions: Preoperative cachexia was associated with higher rates of TO. Higher TO was attributed to patients with overweight or obesity, to a shorter index stay, and more frequent transfers to a secondary facility, but not fewer complications. Cachexia was not associated with worse long-term survival, and a "double jeopardy" between cachexia and receiving NAT was not found.

Background: The ACOSOG Z11102 trial demonstrated the safety of breast-conserving surgery (BCS) with adjuvant radiation in women with multiple ipsilateral breast cancer (MIBC) undergoing upfront surgery, reporting a 5-year local recurrence (LR) of ~3%. However, the oncologic safety of BCS in women with MIBC receiving neoadjuvant systemic therapy (NST) remains uncertain.

Patients and methods: Patients with stage I-III unifocal or MIBC who underwent BCS following NST from 2016 to 2023 were identified in a prospectively maintained institutional database. MIBC was defined preoperatively as the presence of 2-3 foci of biopsy-proven breast cancer with at least 2 cm of intervening normal breast tissue and at least one focus of invasive disease.

Results: A total of 1515 patients were identified: 73 (4.8%) with MIBC and 1442 (95.2%) with unifocal disease. Baseline clinicopathologic characteristics were similar between groups. Median age was 55 years, and most received neoadjuvant chemotherapy (82.2% vs. 83.4%). Molecular subtype distribution was similar between cohorts (p = 0.97). Of the patients with MIBC, 48 (65.8%) underwent single-site lumpectomy, 23 (31.5%) two-site lumpectomy, and 2 (2.7%) three-site lumpectomy. At median follow-up of 34.7 months, there was no difference in LR (1.4% vs. 3.1%, p = 0.40), distant recurrence (5.5% vs. 4.6%, p = 0.37), or breast cancer mortality (4.1% vs. 2.6%, p = 0.45) between groups.

Conclusions: In this retrospective analysis of women with MIBC who underwent BCS after NST, local recurrence (LR) was 1.4% at 3-year median follow-up, which was similar to patients with unifocal breast cancer. These findings suggest BCS is a safe surgical option in well-selected patients with MIBC undergoing NST.

Background: Cryoablation is emerging as a minimally invasive alternative to lumpectomy for select women with early-stage breast cancer. The FROST trial (NCT01992250) was a prospective, phase 2 multicenter study evaluating the outcome of cryoablation in the management of stage I, hormone receptor-positive, human epidermal growth factor receptor 2 (HER2)-negative, node-negative invasive ductal carcinoma.

Methods: Women 50 years old or older with unifocal, ultrasound-visible tumors were stratified by age: stratum 1 (age ≥70 years, endocrine therapy only) and stratum 2 (age 50-69 years, endocrine therapy + radiotherapy + optional sentinel node biopsy). Cryoablation was performed using a single cryoprobe under ultrasound guidance. Core biopsy 6 months after ablation was performed to confirm complete ablation. Patients were followed with clinical exams and imaging.

Results: The study included 83 completed cryoablations and follow-up evaluations. The median tumor size was 9 mm. More than 85% of the subjects in each group received endocrine therapy (stratum 1 [89%, 43/48], stratum 2 [85.7%, 30/35]) and 74.3% (26/35) of the subjects in stratum 2 received recommended whole-breast radiation. Of the 83 patients, 82 received a post-ablation core biopsy 6 months after cryoablation showing no residual cancer, and 1 patient declined a core biopsy. During a median follow-up period of 6.1 years, the 5-year ipsilateral breast tumor recurrence rate (IBTR) was 3.64% overall (stratum 1, 2.08%; stratum 2, 5.80%). The invasive IBTR-free survival rate was 97.59% overall (stratum 1, 97.92%; stratum 2, 97.14%). No serious adverse events occurred.

Conclusions: The FROST trial adds to the growing body of literature supporting the efficacy and safety of cryoablation and supports ongoing research on cryoablation as a strategy for de-escalating breast cancer therapy.

Background: Curative-intent pancreatectomy for pancreatic ductal adenocarcinoma (PDAC) carries high rates of morbidity and recurrence. Patients with operable PDAC face a complex decision of whether to pursue resection. Factors impacting the decision-making process and decisional regret (DR) are relatively unexplored.

Patients and methods: Patients with PDAC who underwent curative-intent pancreatectomy at least 6 months prior to study recruitment completed validated surveys assessing DR, health literacy, shared decision-making, and quality of life.

Results: Overall, 60 patients (48% female) completed all surveys. Postoperative DR (DRS > 1) was reported in 21 (35%) patients. The median time from surgery to survey completion was 55 months in the DR Present group and 33 months in the DR Absent group (P = 0.895). Receipt of systemic therapy, operative characteristics, and postoperative course were similar between groups. The DR Present group had lower rates of obtaining advanced educational degrees (48% versus 76%, P = 0.031), lower health literacy (BRIEF score 14.8 ± 4.2 versus 17.0 ± 2.5, P = 0.014), and greater discordance between preferred and actual roles in the decision making process (Cohen's kappa 0.672 [95% CI 0.530-0.814] versus 0.912 [95% CI 0.852-0.972], P < 0.001). The DR Present group reported worse physical ability (EORTC score 77.8 ± 21.8 versus 91.3 ± 11.8, P = 0.014), functional role (EORTC score 63.5 ± 33.2 versus 85.5 ± 27.6, P = 0.008), and social activity (EORTC score 65.9 ± 35.9 versus 85.9 ± 21.4, P = 0.026) scores.

Conclusions: Lower health literacy and discordance in preferred and actual decision-making roles negatively impacts post-pancreatectomy DR. Adequate preoperative counseling on potential quality of life changes is critical for informed decision-making.