Prediction of Recurrence Pattern of Pancreatic Cancer Post-Pancreatic Surgery Using Histology-Based Supervised Machine Learning Algorithms: A Single-Center Retrospective Study

Preoperative Prediction of Long-Term Survival After Surgery in Patients with Resectable Pancreatic Ductal Adenocarcinoma

BACKGROUND

Although some clinical trials have demonstrated the benefits of neoadjuvant therapy for resectable pancreatic ductal adenocarcinoma (PDAC), its optimal candidate has not been clarified. This study aimed to detect predictive prognostic factors for resectable PDAC patients who underwent upfront surgery and identify patient cohorts with long-term survival without neoadjuvant therapy.

PATIENTS AND METHODS

A total of 232 patients with resectable PDAC who underwent upfront surgery between January 2008 and December 2019 were evaluated.

RESULTS

The median overall survival (OS) time and 5-year OS rate of resectable PDAC with upfront surgery was 31.5 months and 33.3%, respectively. Multivariate analyses identified tumor diameter in computed tomography (CT) ≤ 19 mm [hazard ratio (HR) 0.40, p < 0.001], span-1 within the normal range (HR 0.54, p = 0.023), prognostic nutritional index (PNI) ≥ 44.31 (HR 0.51, p < 0.001), and lymphocyte-to-monocyte ratio (LMR) ≥ 3.79 (HR 0.51, p < 0.001) as prognostic factors that influence favorable prognoses after upfront surgery. According to the prognostic prediction model based on these four factors, patients with four favorable prognostic factors had a better prognosis with a 5-year OS rate of 82.4% compared to others (p < 0.001). These patients had a high R0 resection rate and a low frequency of tumor recurrence after upfront surgery.

CONCLUSIONS

We identified patients with long-term survival after upfront surgery by prognostic prediction model consisting of tumor diameter in CT, span-1, PNI, and LMR. Evaluation of anatomical, biological, nutritional, and inflammatory factors may be valuable to introduce an optimal treatment strategy for resectable PDAC.

Organ-specific variations in tumor marker dynamics in postoperative pancreatic cancer recurrence: Trends in lung and liver recurrence highlighting biological heterogeneity

BACKGROUND/OBJECTIVES

Although tumor recurrence after surgical resection in pancreatic cancer (PC) is generally considered incurable, it is well-accepted that clinical presentations and outcomes vary according to the recurrent sites (e.g., liver vs. lung recurrence), suggesting a possible biological inhomogeneity of PC recurrence. Understanding the behavior of biological factors, specifically tumor markers (TMs), at different recurrence sites may contribute to individualized treatment strategies. Therefore, this study aimed to compare the dynamics of pre-recurrence TMs at liver and lung recurrence sites.

METHODS

Patients with isolated postoperative liver or lung recurrence as their first recurrence were enrolled. Starting from the recurrence date confirmed by imaging examinations, the values of TMs (carbohydrate antigen 19-9: CA19-9; carcinoembryonic antigen: CEA) were retrospectively evaluated 6 and 3 months before recurrence and at the time of recurrence.

RESULTS

Patients with liver recurrence displayed a significant increase in CA19-9 and CEA levels from as early as 6 months before recurrence. Contrastingly, patients with lung recurrence demonstrated a significant elevation of CA19-9 levels starting from 3 months before recurrence, with no increase in CEA levels, even at the time of recurrence. The relative change in CA19-9 and CEA levels during each period were significantly lower in patients with lung recurrence.

CONCLUSIONS

Both TMs exhibited organ-specific variations in patients with postoperative PC recurrence. This disparity may reflect the biological heterogeneity of PC between recurrence patterns, thereby highlighting the importance of conducting postoperative follow-up with consideration of this fact.

Prognostication for recurrence patterns after curative resection for pancreatic ductal adenocarcinoma

Backgrounds/Aims

This study aimed to investigate patterns and factors affecting recurrence after curative resection for pancreatic ductal adenocarcinoma (PDAC).

Methods

Consecutive patients who underwent curative resection for PDAC (2011-21) and consented to data and tissue collection (Barts Pancreas Tissue Bank) were followed up until May 2023. Clinico-pathological variables were analysed using Cox proportional hazards model.

Results

Of 91 people (42 males [46%]; median age, 71 years [range, 43-86 years]) with a median follow-up of 51 months (95% confidence intervals [CIs], 40-61 months), the recurrence rate was 72.5% (n = 66; 12 loco-regional alone, 11 liver alone, 5 lung alone, 3 peritoneal alone, 29 simultaneous loco-regional and distant metastases, and 6 multi-focal distant metastases at first recurrence diagnosis). The median time to recurrence was 8.5 months (95% CI, 6.6-10.5 months). Median survival after recurrence was 5.8 months (95% CI, 4.2-7.3 months). Stratification by recurrence location revealed significant differences in time to recurrence between loco-regional only recurrence (median, 13.6 months; 95% CI, 11.7-15.5 months) and simultaneous loco-regional with distant recurrence (median, 7.5 months; 95% CI, 4.6-10.4 months; p = 0.02, pairwise log-rank test). Significant predictors for recurrence were systemic inflammation index (SII) ≥ 500 (hazard ratio [HR], 4.5; 95% CI, 1.4-14.3), lymph node ratio ≥ 0.33 (HR, 2.8; 95% CI, 1.4-5.8), and adjuvant chemotherapy (HR, 0.4; 95% CI, 0.2-0.7).

Conclusions

Timing to loco-regional only recurrence was significantly longer than simultaneous loco-regional with distant recurrence. Significant predictors for recurrence were SII, lymph node ration, and adjuvant chemotherapy.

Machine Learning of Histopathological Images Predicts Recurrences of Resected Pancreatic Ductal Adenocarcinoma With Adjuvant Treatment

OBJECTIVES

Pancreatic ductal adenocarcinoma is an intractable disease with frequent recurrence after resection and adjuvant therapy. The present study aimed to clarify whether artificial intelligence-assisted analysis of histopathological images can predict recurrence in patients with pancreatic ductal adenocarcinoma who underwent resection and adjuvant chemotherapy with tegafur/5-chloro-2,4-dihydroxypyridine/potassium oxonate.

MATERIALS AND METHODS

Eighty-nine patients were enrolled in the study. Machine-learning algorithms were applied to 10-billion-scale pixel data of whole-slide histopathological images to generate key features using multiple deep autoencoders. Areas under the curve were calculated from receiver operating characteristic curves using a support vector machine with key features alone and by combining with clinical data (age and carbohydrate antigen 19-9 and carcinoembryonic antigen levels) for predicting recurrence. Supervised learning with pathological annotations was conducted to determine the significant features for predicting recurrence.

RESULTS

Areas under the curves obtained were 0.73 (95% confidence interval, 0.59-0.87) by the histopathological data analysis and 0.84 (95% confidence interval, 0.73-0.94) by the combinatorial analysis of histopathological data and clinical data. Supervised learning model demonstrated that poor tumor differentiation was significantly associated with recurrence.

CONCLUSIONS

Results indicate that machine learning with the integration of artificial intelligence-driven evaluation of histopathological images and conventional clinical data provides relevant prognostic information for patients with pancreatic ductal adenocarcinoma.

Clinical benefits of pulmonary resection for lung metastases from pancreatic cancer

PURPOSE

Systemic chemotherapy is generally used for metastatic pancreatic cancer; however, pulmonary resection may be a treatment option for lung oligometastases from pancreatic cancer. The current study aimed to clarify the oncological outcomes and clinical benefits of pulmonary resection for lung metastases.

METHODS

Of 510 patients who underwent pancreatic resection for pancreatic cancer, 44 patients with recurrence of isolated lung metastases and one patient with simultaneous lung metastases were evaluated.

RESULTS

Of the 45 patients, 20 patients were selected as candidates for pulmonary resection based on clinical factors such as recurrence-free interval (RFI) from pancreatectomy to lung metastases, number of lung metastases, and serum CA19-9 level. The post-recurrent survival of patients with pulmonary resection was significantly better than that of patients without pulmonary resection. Fourteen of the 20 patients with pulmonary resection developed tumor recurrence with a median disease-free survival (DFS) of 15 months. Univariate analyses revealed that an RFI from pancreatectomy to lung metastases of ≥28 months was associated with better DFS after pulmonary resection. Of the 14 patients with an RFI of ≥28 months, pulmonary resection resulted in prolonged chemotherapy-free interval in 12 patients. Furthermore, repeat pulmonary resection for recurrent tumors after pulmonary resection led to further cancer-free interval in some cases.

CONCLUSIONS

Although many patients had tumor recurrence after pulmonary resection, pulmonary resection for lung metastases from pancreatic cancer may provide prolonged cancer-free interval without the need for chemotherapy. Pulmonary resection should be performed for the patients with a long RFI from pancreatectomy to lung metastases.

New approach of circumferential lymph node dissection around the superior mesenteric artery for pancreatic cancer during pancreaticoduodenectomy (with video)

PURPOSE

Various approaches have been reported for the resection of the nervous and lymphatic tissues around the superior mesenteric artery (SMA) during pancreaticoduodenectomy (PD) for pancreatic cancer. We developed a new procedure for circumferential lymph node dissection around the SMA to minimize local recurrence.

METHODS

We included 24 patients who underwent PD with circumferential lymph node dissection around the SMA (circumferential dissection) and 94 patients who underwent classical mesopancreatic dissection (classical dissection) between 2019 and 2021. The technical details of this new method are described in the figures and videos, and the clinical characteristics and outcomes of this technique were compared with those of classical dissection.

RESULTS

The median follow-up durations in the circumferential and classical dissection groups were 39 and 36 months, respectively. The patients' characteristics, including tumor resectability, preoperative and adjuvant chemotherapy rates, postoperative complication rates, and tumor stage, were similar between the two groups. No differences were observed in recurrence-free survival and overall survival between the two groups; however, the classical dissection group tended to have more local recurrences than the circumferential dissection group (8.3% vs. 33.3%, P = 0.168). Although no case of nodular-type recurrence after circumferential dissection was observed, 61.1% of local recurrences after classical dissection were of the nodular-type, and 36.4% were located on the left side of the SMA.

CONCLUSIONS

Performing circumferential lymph node dissection around the SMA during PD can be conducted safely with minimal risks of local recurrence and may enhance the completeness of local resection.

Machine learning in pancreas surgery, what is new? literature review

Background

Machine learning (ML) is an inquiry domain that aims to establish methodologies that leverage information to enhance performance of various applications. In the healthcare domain, the ML concept has gained prominence over the years. As a result, the adoption of ML algorithms has become expansive. The aim of this scoping review is to evaluate the application of ML in pancreatic surgery.

Methods

We integrated the preferred reporting items for systematic reviews and meta-analyses for scoping reviews. Articles that contained relevant data specializing in ML in pancreas surgery were included.

Results

A search of the following four databases PubMed, Cochrane, EMBASE, and IEEE and files adopted from Google and Google Scholar was 21. The main features of included studies revolved around the year of publication, the country, and the type of article. Additionally, all the included articles were published within January 2019 to May 2022.

Conclusion

The integration of ML in pancreas surgery has gained much attention in previous years. The outcomes derived from this study indicate an extensive literature gap on the topic despite efforts by various researchers. Hence, future studies exploring how pancreas surgeons can apply different learning algorithms to perform essential practices may ultimately improve patient outcomes.

Prediction of early-stage melanoma recurrence using clinical and histopathologic features

Prognostic analysis for early-stage (stage I/II) melanomas is of paramount importance for customized surveillance and treatment plans. Since immune checkpoint inhibitors have recently been approved for stage IIB and IIC melanomas, prognostic tools to identify patients at high risk of recurrence have become even more critical. This study aims to assess the effectiveness of machine-learning algorithms in predicting melanoma recurrence using clinical and histopathologic features from Electronic Health Records (EHRs). We collected 1720 early-stage melanomas: 1172 from the Mass General Brigham healthcare system (MGB) and 548 from the Dana-Farber Cancer Institute (DFCI). We extracted 36 clinicopathologic features and used them to predict the recurrence risk with supervised machine-learning algorithms. Models were evaluated internally and externally: (1) five-fold cross-validation of the MGB cohort; (2) the MGB cohort for training and the DFCI cohort for testing independently. In the internal and external validations, respectively, we achieved a recurrence classification performance of AUC: 0.845 and 0.812, and a time-to-event prediction performance of time-dependent AUC: 0.853 and 0.820. Breslow tumor thickness and mitotic rate were identified as the most predictive features. Our results suggest that machine-learning algorithms can extract predictive signals from clinicopathologic features for early-stage melanoma recurrence prediction, which will enable the identification of patients that may benefit from adjuvant immunotherapy.