Predictors Of Disease-specific Survival Research Articles

Epidemiological characteristics and clinical course of eyelid squamous cell carcinoma patients from a large tertiary centre between 2009 and 2020

Background/aimsTo assess epidemiological tumour features, risk factors, clinical management and outcome of eyelid squamous cell carcinoma (SCC) and changes thereof. Furthermore, we searched for validating predictors of the American Joint...

Young age is not a predictor of disease specific survival in oral cancer: A multi-institutional study

BackgroundOver the last few decades evidence has accumulated for increasing incidence of oral cavity squamous cell carcinoma (OSCC) in a younger cohort. Prior studies examining the effect of age at diagnosis on prognosis have produced conflicting data. MethodsA multi-institutional cohort study was performed across 6 different sites in Australia, Canada, India and Singapore. Disease-free (DFS), overall (OS) and disease-specific (DSS) survival were analysed. The association of the number of adverse features with survival outcomes was investigated. ResultsFrom 3179 patients, age was a significant predictor of OS with patients older than 45 years having a 66% increased risk of death (HR 1.66, 95%CI 1.33 – 2.07, p < 0.001). The number of adverse features was a significant predictor of OS with 3 or more adverse features having a 199% increased risk (HR 2.99, 95%CI 2.61–3.43. p < 0.001). The estimate effect was greater in patients ≤ 45 years (HR 3.49 vs HR 2.81). Age was not a significant predictor of DSS with similar rates of death from OSCC in multivariable models. The number of adverse features was a significant predictor of DFS with ≥ 3 adverse features having a 140% increased risk of death. The number of adverse features was a significant predictor of DSS with ≥ 3 adverse features having a 230% increased risk of disease specific death. ConclusionsAge is not an independent predictor of disease specific mortality in OSCC. Differences in outcomes are due to the confounding effect of adverse clinicopathological features and the ability to tolerate surgery and adjuvant therapy.

Automated digital TIL analysis (ADTA) adds prognostic value to standard assessment of depth and ulceration in primary melanoma

Accurate prognostic biomarkers in early-stage melanoma are urgently needed to stratify patients for clinical trials of adjuvant therapy. We applied a previously developed open source deep learning algorithm to detect tumor-infiltrating lymphocytes (TILs) in hematoxylin and eosin (H&E) images of early-stage melanomas. We tested whether automated digital (TIL) analysis (ADTA) improved accuracy of prediction of disease specific survival (DSS) based on current pathology standards. ADTA was applied to a training cohort (n = 80) and a cutoff value was defined based on a Receiver Operating Curve. ADTA was then applied to a validation cohort (n = 145) and the previously determined cutoff value was used to stratify high and low risk patients, as demonstrated by Kaplan–Meier analysis (p ≤ 0.001). Multivariable Cox proportional hazards analysis was performed using ADTA, depth, and ulceration as co-variables and showed that ADTA contributed to DSS prediction (HR: 4.18, CI 1.51–11.58, p = 0.006). ADTA provides an effective and attainable assessment of TILs and should be further evaluated in larger studies for inclusion in staging algorithms.

Impact of lymph node staging systems in predicting outcome in patients with ampullary cancer.

Backgrounds/AimsLymph node (LN) metastasis though, is a poor prognostic factor for ampullary carcinoma (APC), the impact of Lymph node ratio (LNR) and Logarithm odds of positive lymph node (LODDS) in the long-term survival remains controversial. We evaluated the factors affecting the long-term outcome in APC patients with emphasis on LNR and LODDS.MethodsThe prospectively collected data of 198 patients who underwent pancreatoduodenectomy for APC was analyzed after excluding 12 patients for various reasons. Factors affecting Disease specific survival (DSS) and Recurrence free survival (RFS) were analyzed with special reference to LN positivity, LNR and LODDS.ResultsOut of 186, 117 (62.9%) patients were alive at a median follow-up of 39.5 months and 72 (38.7%) developed recurrence. The overall 5-year DSS was 59.3% & RFS 54.9%. Univariate analysis showed T-stage, tumor differentiation, perineural invasion, LN positivity, LNR and LODDS was significantly affected DSS and RFS. On multivariate analysis, perineural invasion, LN positivity, LNR and LODDS lost its significance for DSS and RFS. AUC for prediction of DSS and RFS for LNR was 0.654 (p<0.001) & 0.629 (p=0.003) respectively and for LODDS, it was 0.697 (p<0.001) & 0.677 (p=0.001) respectively. Sensitivity and specificity of LNR (0.1) for DSS were 37.7% & 83.8% and for RFS were 36.1% & 83.3%; for LODDS (−1.00), sensitivity and specificity for DSS was 62.3% and 67.5% and for RFS it was 59.7% and 66.7% respectively.ConclusionsLNR and LODDS although independently seem to affect the RFS and DSS, albeit have a low sensitivity and specificity in predicting DSS and RFS.

Prognostic Significance of Oxidation Pathway Mutations in Recurrent Laryngeal Squamous Cell Carcinoma.

Simple SummaryOrgan preservation protocols have become first line therapy for the majority of advanced laryngeal cancers. Unfortunately, up to one third of patients will develop recurrent disease requiring salvage surgery. These tumors tend to display aggressive features when compared to primary disease. The aim of this study is to identify genomic alterations associated with poor prognosis in the recurrent setting to guide precision therapy and identify potential targetable pathways. Here we show that mutations in the oxidation pathway, specifically the KEAP1-NFR2 pathway, predict survival in a cohort of patients undergoing salvage laryngectomy.Organ preservation protocols are commonly used as first line therapy for advanced laryngeal cancer. Recurrence thereafter is associated with poor survival. The aim of this study is to identify genetic alterations associated with survival among patients with recurrent laryngeal cancer undergoing salvage laryngectomy. Sixty-two patients were sequenced using a targeted panel, of which twenty-two also underwent transcriptome sequencing. Alterations were grouped based on biologic pathways and survival outcomes were assessed using Kaplan-Meier analysis and multivariate cox regression. Select pathways were evaluated against The Cancer Genome Atlas (TCGA) data. Patients with mutations in the Oxidation pathway had significantly worse five-year disease specific survival (1% vs. 76%, p = 0.02), while mutations in the HN-Immunity pathway were associated with improved five-year disease specific survival (100% vs. 62%, p = 0.02). Multivariate analysis showed mutations in the Oxidation pathway remained an independent predictor of disease specific survival (HR 3.2, 95% CI 1.1–9.2, p = 0.03). Transcriptome analysis of recurrent tumors demonstrated that alterations in the Oxidation pathway were associated a positive Ragnum hypoxia signature score, consistent with enhanced pathway activity. Further, TCGA analyses demonstrated the prognostic value of oxidation pathway alterations in previously untreated disease. Alterations in the Oxidation pathway are associated with survival among patients with recurrent laryngeal cancer. These prognostic genetic biomarkers may inform precision medicine protocols and identify putatively targetable pathways to improve survival in this cohort.

Recurrent retroperitoneal sarcomas: Clinical outcomes of surgical treatment and prognostic factors

Background and purposeLocoregional recurrence after resection of primary retroperitoneal sarcoma (RPS) is a challenging therapeutic issue. The objective of this study was to identify clinicopathological factors predictive of overall survival (OS) and disease specific survival (DSS) after reoperation for recurrent RPS. Patients and methodsWe retrospectively collected data from the medical records of 800 patients who underwent resection for sarcoma at our Institution, from 1983 to 2015. Among these patients, 120 were treated for retroperitoneal sarcoma and 55 had a locoregional recurrence (LR). Four of them did not undergo surgery and thus were excluded from this study leaving 51 cases available for data analysis. Univariate and multivariate survival analyses were performed to identify prognostic factors. ResultsMedian overall survival was 33 months. The 1-year, 3-year and 5-year OS rates were 75.5%, 47.1% and 31.6% respectively. Multivariate Cox regression analysis suggested that extension of surgery (P = 0.026), surgical margin status (P = 0.015) and histological grade of recurrent tumor (P = 0.047) were independent prognostic factors for OS. Median DSS was 48 months. The 1-year, 3-year and 5-year DSS rates were 79.2%, 53.1% and 40.9%, respectively. At multivariate analysis, predictors of DSS were extension of surgery (P = 0.004), margin status (P = 0.011), histological grade of recurrent tumor (P = 0.008), and disease free interval (DFI) (P = 0.020). As regards histological subtype of recurrent RPS, at univariate analysis, well-differentiated liposarcoma (WDLS) was associated with better OS and DSS (P = 0.052 and P = 0.016 respectively) compared to dedifferentiated liposarcoma (DDLS). ConclusionsAccording to our findings, surgery is more beneficial in patients with low-grade sarcoma, WDLS and long DFI. The achievement of clear resection margins, rather than performing a multivisceral resection, appears to be a key factor to improve OS and DSS.

Laparoscopic liver resection versus percutaneous radiofrequency ablation for small hepatocellular carcinoma

BackgroundLaparoscopic liver resection (LLR) and radiofrequency ablation (RFA) play central roles to treat early-stage hepatocellular carcinoma (HCC, ≤3 cm, 1–3 nodules, and no macrovascular involvement), although data are lacking regarding whether LLR or RFA is preferable. This study aimed to compare outcomes of both treatments for small HCCs. MethodsTreatment outcomes of small HCCs were compared between all the minor LLRs performed between 2005 and 2016 and RFAs performed between 2011 and 2016 at Kyoto University. ResultsA total of 85 and 136 patients underwent LLR and RFA, respectively. Patients that underwent LLR had higher incidence of blood transfusions, complications, and longer hospital stay. Overall and disease-specific survival rates were similar between LLR and RFA; however, recurrence-free (49.2% vs. 22.1% at 3-year) and local recurrence-free survival rates (94.9% vs. 63.6% at 3-year) were higher after LLR. Multivariate analyses identified that multiple nodules and 65-year-old and above are predictors of disease-specific survival, and that RFA is a predictor of recurrence and local recurrence. ConclusionRFA is less invasive, although both LLR and RFA are safe and effective. LLR provides better local control with superior recurrence-free and local-recurrence free survival. These results help optimize treatment selection based on patient-specific factors.

Abstract 2096: A deep learning system to predict disease-specific survival in stage II and stage III colorectal cancer

Abstract Accurate prognosis in colorectal cancer can have important implications for clinical management. Here, we develop a deep learning system (DLS) to first identify invasive cancer and then directly predict disease specific survival (DSS) for stage II and stage III colorectal cancer using only digitized histopathology whole-slide images. The DLS was trained using slides from 1173 stage II and 1266 stage III cases (18,304 total slides) and was evaluated on a held-out test set of 601 stage II and 638 stage III cases (9,340 total slides). The area under the receiver operating characteristic curve (AUC) for 5-year DSS prediction was 68.0 for stage II (95% CI 62.2-73.1) and 65.5 for stage III (95% CI 61.1-70.0). For stage II, 5-year DSS was 64% for DLS-predicted high-risk cases versus 89% for DLS-predicted low-risk cases (upper and lower risk quartiles; p&amp;lt;0.001, log rank test). For stage III, 5-year DSS was 35% for DLS-predicted high-risk cases versus 66% for DLS-predicted low-risk cases (upper and lower risk quartiles; p&amp;lt;0.001, log rank test). In a multivariable Cox model, the DLS prediction remained significantly associated with DSS after adjusting for T-category, N-category, age, gender, tumor grade, and lymphovascular invasion (stage II: adjusted hazard ratio 1.55, 95% CI 1.33-1.81, p&amp;lt;0.0001; stage III: adjusted hazard ratio 1.35, 95% CI (1.21-1.51), p&amp;lt;0.0001). Finally, a combined proportional-hazards model using the DLS along with baseline clinicopathologic information provided better risk prediction than the DLS or baseline information alone, increasing 5-year AUC over the baseline-only model by 8.9 points (95% CI 3.9-13.6) and 5.3 points (95% CI 2.3-8.4) for stages II and III, respectively. Taken together, these findings demonstrate that the DLS provides significant prognostic value and risk stratification in both stage II and stage III colorectal cancer, and can be combined with known risk features to further improve prognostic accuracy. This represents novel work to train a DLS to directly predict patient outcomes using whole-slide images and weakly supervised learning. The ability to use non-annotated slides as input has important implications for possible clinical applications and the features learned by the model may also help to identify new prognosis-associated morphologic factors in colorectal cancer. Additional work is ongoing to confirm the utility of these findings, such as validation in additional datasets and interpretability experiments to better understand the features learned by the DLS for these predictions. Citation Format: Ellery Wulczyn, David F. Steiner, Melissa Moran, Markus Plass, Robert Reihs, Heimo Mueller, Apaar Sadhwani, Yuannan Cai, Isabelle Flament, Po-Hsuan Cameron Chen, Yun Liu, Martin C. Stumpe, Zhaoyang Xu, Kurt Zatloukal, Craig H. Mermel. A deep learning system to predict disease-specific survival in stage II and stage III colorectal cancer [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 2096.

Management of adenoid cystic carcinoma of the head and neck: a single-institute study with over 25-year follow-up

BackgroundAdenoid cystic carcinoma is a rare malignant tumor arising from exocrine glands such as the major and minor salivary glands of the paranasal sinuses or the external auditory canal. Although multiple retrospective clinical studies of ACC have been reported to date, clinical questions, such as 1) long-term prognosis beyond 20 years, 2) usefulness and suitability for treatment of therapeutic interventions, 3) therapeutic goal to aim for, and 4) prognosis by recurrence sites, are still unclear.MethodsTo improve understanding and management of adenoid cystic carcinoma of the head and neck (ACC), a retrospective study with 58 new ACC cases between 1991 and 2016 was performed. The median observation period was 66.8 months (range 3–316 months). The overall clinical stages were as follows: I, 6.9%; II, 25.9%; III, 19.0%; and IV, 48.2%. Histology was cribriform/tubular type (C-T type) in 62.0% and solid type in 27.5%. The main treatment strategy was definitive surgery, which was performed in 75.2% of cases.ResultsOverall 10-year, 20-year, and 25-year survivals were 63.7, 27.3, and 20.0%, respectively. Similarly, disease-specific survival (DSSs) was 65.7, 51.2, and 38.4%, respectively, and disease-free survival was 25.2, 9.4, and 9.4%, respectively. Conducting surgery (HR: 0.19, 95% CI: 0.06–0.61, p = 0.005) and C-T type (HR: 0.32, 95% CI: 0.11–0.93, p = 0.036) were independent prognostic predictors of DSS. DSS was significantly prolonged after salvage surgery for both locoregional recurrence (p = 0.004) and lung metastatic recurrence (p = 0.012, vs best supportive care).ConclusionsIn ACC cases, both initial surgical treatment and repetitive surgical resection of resectable recurrent lesions, including both locoregional and lung metastases, resulted in longer survival. The major goal of treatment for ACC may be long-term survival including cancer-bearing survival, resulting in either natural death or intercurrent-disease death, since judging cure of ACC is almost impossible.Trial registrationRetrospectively registered.

The Prognostic Factors of Preoperative Prognostic Nutritional Index and Radiological Findings of Solid Pseudopapillary Tumors of Pancreas: A Single-Center Experience of 14 Years.

IntroductionSolid pseudopapillary tumors of pancreas (SPTP) is an indolent rare tumor with malignant potential. The prediction of malignancy is an enigma. The aim of this study is to explore the relationship between operative measurements and malignancy prognosis of SPTP patients.MethodsA cohort of consecutive 102 patients were enrolled in this study. Preoperative measurements and clinical outcomes were analyzed.ResultsEighteen patients (17.6%) were confirmed as malignant. The malignant SPTP of the optimal cut-off value was 47.9 (p=0.012) for prognostic nutritional index (PNI). The value of PNI≤47.9 and incomplete capsule were significantly correlated with malignancy. Univariate analysis showed that the PNI≤47.9 (p=0.013) and incomplete capsule (p<0.001) were predictors of disease-specific survival (DSS). Multivariate analysis identified the PNI≤47.9 (p=0.036) and incomplete capsule (p=0.023) as the independent prognostic factors of DSS. The new score of 0,1,2 based on PNI and capsule presence stratified the patients into 3 groups. The patients with low PNI and incomplete capsule achieved the worst prognosis.ConclusionThe combination test of operative PNI and capsule presence would be a reliable indicator of the aggressive natural history of SPTP.

Deep learning-based survival prediction for multiple cancer types using histopathology images.

Providing prognostic information at the time of cancer diagnosis has important implications for treatment and monitoring. Although cancer staging, histopathological assessment, molecular features, and clinical variables can provide useful prognostic insights, improving risk stratification remains an active research area. We developed a deep learning system (DLS) to predict disease specific survival across 10 cancer types from The Cancer Genome Atlas (TCGA). We used a weakly-supervised approach without pixel-level annotations, and tested three different survival loss functions. The DLS was developed using 9,086 slides from 3,664 cases and evaluated using 3,009 slides from 1,216 cases. In multivariable Cox regression analysis of the combined cohort including all 10 cancers, the DLS was significantly associated with disease specific survival (hazard ratio of 1.58, 95% CI 1.28–1.70, p<0.0001) after adjusting for cancer type, stage, age, and sex. In a per-cancer adjusted subanalysis, the DLS remained a significant predictor of survival in 5 of 10 cancer types. Compared to a baseline model including stage, age, and sex, the c-index of the model demonstrated an absolute 3.7% improvement (95% CI 1.0–6.5) in the combined cohort. Additionally, our models stratified patients within individual cancer stages, particularly stage II (p = 0.025) and stage III (p<0.001). By developing and evaluating prognostic models across multiple cancer types, this work represents one of the most comprehensive studies exploring the direct prediction of clinical outcomes using deep learning and histopathology images. Our analysis demonstrates the potential for this approach to provide significant prognostic information in multiple cancer types, and even within specific pathologic stages. However, given the relatively small number of cases and observed clinical events for a deep learning task of this type, we observed wide confidence intervals for model performance, thus highlighting that future work will benefit from larger datasets assembled for the purposes for survival modeling.

Neutrophil-lymphocyte ratio associated with poor prognosis in oral cancer: a retrospective study

BackgroundPrognostic biomarkers provide essential information about a patient’s overall outcome. However, existing biomarkers are limited in terms of either sample collection, such as requiring tissue specimens, or the process, such as prolonged time for analysis. In view of the need for convenient and non-invasive prognostic biomarkers for oral cancer, we aimed to investigate the prognostic values of neutrophil-to-lymphocyte ratio, lymphocyte-to-monocyte ratio, and platelet-to-lymphocyte ratio in patient survival. We also aimed to explore the associations of these ratios with the clinicopathologic characteristics of Japanese oral squamous cell carcinoma patients.MethodsThis study was a non-randomized retrospective cohort study in a tertiary referral center. We included 433 patients (246 men, 187 women) who underwent radical surgery for oral cancers between January 2001 and December 2013. We evaluated various risk factors for poor prognosis including neutrophil-to-lymphocyte ratio, lymphocyte-to-monocyte ratio, and platelet-to-lymphocyte ratio with univariate and multivariate analyses. The disease-specific survival and overall survival rates of patients were compared among the factors and biomarkers.ResultsIn multivariable Cox proportional hazards analysis, high neutrophil-to-lymphocyte ratio (hazard ratio 2.87, 95% confidence interval 1.59–5.19, P < 0.001), moderately or poorly differentiated histology (hazard ratio 2.37, 95% confidence interval 1.32–4.25, P < 0.001), and extranodal extension (hazard ratio 1.95, 95% confidence interval 1.13–3.35, P = 0.016) were independent predictors of disease-specific survival. High neutrophil-to-lymphocyte ratio (hazard ratio 2.30, 95% confidence interval 1.42–3.72, P < 0.001), moderately or poorly differentiated (hazard ratio 1.72, 95% confidence interval 1.07–2.76, P = 0.025), and extranodal extension (hazard ratio 1.79, 95% confidence interval 1.13–2.84, P = 0.013) were independent predictors of overall survival.ConclusionsNeutrophil-to-lymphocyte ratio might be a potential independent prognostic factor in Japanese oral squamous cell carcinoma patients.

Long-term oncological safety of sentinel lymph node biopsy in early-stage cervical cancer.

6006 Background: The goal of this study was to assess disease-free survival (DFS) and disease-specific survival (DSS) in patients with early-stage cervical cancer who underwent bilateral sentinel lymph node (BSLN) biopsy alone versus bilateral pelvic lymphadenectomy (BPL). Methods: An ancillary analysis of two prospective multicentric trials on SLN biopsy for cervical cancer (SENTICOL I and II) was performed. All patients with early stage cervical cancer (IA to IIB FIGO stage), negative SLN after ultrastaging and negative non-SLN after final pathologic examination were included. Risk-factors of recurrency and disease-specific deaths were determined by Cox proportional hazard models. Kaplan-Meier survival curves were compared by applying log-rank test. Results: Between January 2005 and July 2012, 259 patients met the inclusion criteria: 85 patients underwent only bilateral SLN biopsy whereas 174 patients underwent BPL. None had positive SLN at ultrastaging or positive non-SLN at final pathologic examination. Between the both groups, there was no differences in histology, final FIGO stage and type of surgical approach. In the BPL group, patients had more frequently tumor size larger than 20 mm (22.9% vs 10.7%, p = 0.02) and postoperative radiochemotherapy (10.7% vs 1.6%, p = 0.01). The median follow-up was 47 months (4-127). During the follow-up, 21 patients (8.1%) experienced reccurencies, including 4 nodal recurrences (1.9%), and 9 patients (3.5%) died of cervical cancer. The 5-year DFS and the DSS were similar between BSLN and BPL groups, 94.1% vs 97.7%, p = 0.14 and 88.2% vs 93.7%, p = 0.14 respectively. After controlling for final FIGO stage and margin status, BSLN compared to BPL was not associated with DFS (HR = 1.76, 95%CI = [0.69 – 4.53], p = 0.24) and DSS (HR = 2.5, 95%CI = [0.64 – 9.83], p = 0.19). Only final FIGO stage was independent predictor of DSS. Conclusions: SLN biopsy alone is oncologically safe in early-stage cervical cancer. Full lymphadenectomy could be omitted in case of bilateral negative SLN. Worse prognosis was associated with higher FIGO stage disease.

Cervical Lymph Node Metastases in Oral Squamous Cell Carcinoma-How Much Imaging Do We Need?

Cervical lymph node metastases in oral squamous cell carcinoma (OSCC) are key predictors of disease specific survival. It was therefore the aim of this study to evaluate how much imaging is minimally needed for reliable and efficient identification of cervical lymph node metastases. In this retrospective cross-sectional study, results (metastasis yes/no) of ultrasound (US), computed tomography (CT), magnetic resonance imaging (MRI), and positron emission tomography (PET) were compared to the final histopathological results of the corresponding neck dissection (ND) specimens (metastasis yes/no). A score was calculated to account for cervical lymph node size, shape, clustering, peripheral enhancement, hilus sign, architecture, blood flow, and central necrosis. Sensitivity and specificity were analyzed for each imaging technique separately. In 164 patients diagnosed with OSCC, 96 underwent uni- or bilateral ND (122 ND in total). One hundred percent sensitivity was achieved by CT+MRI, MRI+PET, US+CT+MRI, US+MRI+PET, CT+MRI+PET, and US+CT+MRI+PET. The highest specificity was realized by US with 79% (95% CI [0.698–0.890]). Specificity for CT+MRI and PET+MRI was 51% (95% CI [0.359–0.665]) and 70% (95% CI [0.416–0.984]), respectively. Regarding 100% sensitivity with acceptable specificity, the combination of CT+MRI or PET+MRI appeared to be suitable for staging cervical lymph nodes in primary OSCC.

Modification of the Tumor-Node-Metastasis Staging System for Differentiated Thyroid Carcinoma by Considering Extra-Thyroidal Extension and Lateral Cervical Lymph Node Metastasis.

BackgroundConcerns have arisen about the classification of extra-thyroidal extension (ETE) and lateral cervical lymph node metastasis (N1b) in the 8th edition of the tumor-node-metastasis staging system (TNM-8). This study evaluated the prognostic validity of a modified-TNM staging system, focusing on ETE and N1b, in differentiated thyroid carcinoma (DTC) patients.MethodsThis multicenter retrospective cohort study included 4,878 DTC patients from five tertiary hospitals. In the modified-TNM, T3b in TNM-8 was down-staged to T2, and stage II was subdivided into stages IIA and IIB. Older patients with N1b were reclassified as stage IIB.ResultsThe modified-TNM resulted in staging migration in 540 patients (11%) classified as stage II according to the TNM-8, with 75 (14%), 381 (71%), and 84 patients (16%) classified as stages I, IIA, and IIB, respectively. The 10-year disease-specific survival (DSS) rates in patients classified as stages I, II, III, and IV by TNM-8 were 99.8%, 95.9%, 81.0%, and 41.6%, respectively. The DSS rates of patients classified as stages I, IIA, IIB, III, and IV according to the modified-TNM were 99.8%, 96.4%, 93.3%, 81.0%, and 41.6%, respectively. DSS curves between stages on TNM-8 (P<0.001) and modified-TNM (P<0.001) differed significantly, but the modified-TNM discriminated better than TNM-8. The proportions of variation explained values of TNM-8 and modified-TNM were 6.3% and 6.5%, respectively.ConclusionModification of the TNM staging system focusing on ETE and N1b could improve the prediction of DSS in patients with DTC. Further researches are needed to validate the prognostic accuracy of this modified-TNM staging system.

Deep Learning Based on Standard H&E Images of Primary Melanoma Tumors Identifies Patients at Risk for Visceral Recurrence and Death.

Biomarkers for disease-specific survival (DSS) in early-stage melanoma are needed to select patients for adjuvant immunotherapy and accelerate clinical trial design. We present a pathology-based computational method using a deep neural network architecture for DSS prediction. The model was trained on 108 patients from four institutions and tested on 104 patients from Yale School of Medicine (YSM, New Haven, CT). A receiver operating characteristic (ROC) curve was generated on the basis of vote aggregation of individual image sequences, an optimized cutoff was selected, and the computational model was tested on a third independent population of 51 patients from Geisinger Health Systems (GHS). Area under the curve (AUC) in the YSM patients was 0.905 (P < 0.0001). AUC in the GHS patients was 0.880 (P < 0.0001). Using the cutoff selected in the YSM cohort, the computational model predicted DSS in the GHS cohort based on Kaplan-Meier (KM) analysis (P < 0.0001). The novel method presented is applicable to digital images, obviating the need for sample shipment and manipulation and representing a practical advance over current genetic and IHC-based methods.

Survival study of patients with stage N1-3 testicular seminoma

Objective To explore the independent predictors for disease-specific survival (DSS) rate in patients with stage N1-3 testicular seminoma (TS), and establish a nomogram to predict individual 5-year DSS. Methods The data of N1-3 TS patients registered in the SEER database of National Cancer Institute (USA) from January 2004 to December 2015 were retrospectively analyzed. The 5-year overall survival (OS) rate and DSS rate were calculated using Kaplan-Meier method and the differences among different subgroups were assessed using log-rank test. Besides, the independent predictors of DSS were defined using multivariate Cox regression analysis, and nomogram was drawn using R software. Furthermore, the predictive performance of the nomogram was internally validated using the C-index and calibration plot. Results TNM stage ⅢA (HR=5.604, 95%CI: 1.252-25.083, P=0.024), ⅢB (HR=6.710, 95%CI: 1.923-23.410, P=0.003) and ⅢC (HR=13.189, 95%CI: 3.916-44.420, P<0.001), age at diagnosis ≥45 years old (HR=3.575, 95%CI: 2.014-6.344, P<0.001), and patients without spouse (HR=2.346, 95%CI: 1.406-3.914, P=0.001) were identified as independent risk factors for DSS. On internal validation, the predictive accuracy of our nomogram was 0.751 (C-index: 0.751, 95%CI: 0.694-0.808). Besides, the calibration plot showed that the predicted survival outcomes were highly consistent with the actual survival outcomes. Conclusion The study confirms that age at diagnosis ≥45 years old, TNM stage ≥ⅢA and patients without spouse are the independent risk factors for DSS in TS patients with stage N1-3, and the nomogram for predicting individual 5-year DSS is established. Key words: Seminoma; Lymphatic metastasis; Prognosis; Nomogram

External validation of a postoperative nomogram for the prediction of disease-specific survival in patients with papillary renal cell carcinoma using a large multicenter database.

Based on data retrieved from a comprehensive multicenter database, we externally validated a published postoperative nomogram for the prediction of disease-specific survival (DSS) in patients with papillary renal cell carcinoma (papRCC). A multicenter database containing data of 2325 patients with surgically treated papRCC was used as validation cohort. After exclusion of patients with missing data and patients included in the development cohort, 1372 patients were included in the final analysis. DSS-probabilities according to the nomogram were calculated and compared to actual DSS-probabilities. Subsequently, calibration plots and decision curve analyses were applied. The median follow-up was 38months (IQR 11.8-80.7). Median DSS was not reached. The c-index of the nomogram was 0.71 (95% CI 0.60-0.83). A sensitivity analysis including only patients operated after 1998 delivered a c-index of 0.84 (95% CI 0.77-0.92). Calibration plots showed slight underestimation of nomogram-predicted DSS in probability ranges below 90%: median nomogram-predicted 5-year DSS in the range below 90% was 55% (IQR 20-80), but the median actual 5-year DSS in the same group was 58% (95% CI 52-65). Decision-curve analysis showed a positive net-benefit for probability ranges between a DSS probability of 5% and 85%. The nomogram performance was satisfactory for almost all DSS probabilities; hence it can be recommended for application in clinical routine and for counseling of patients with papRCC.

Clinical characteristics and disease-specific prognostic nomogram for primary gliosarcoma: a SEER population-based analysis

Because the study population with gliosarcoma (GSM) is limited, the understanding of this disease is insufficient. In this study, the authors aimed to determine the clinical characteristics and independent prognostic factors influencing the prognosis of GSM patients and to develop a nomogram to predict the prognosis of GSM patients after craniotomy. A total of 498 patients diagnosed with primary GSM between 2004 and 2015 were extracted from the 18 Registries Research Data of the Surveillance, Epidemiology, and End Results (SEER) database. The median disease-specific survival (DSS) was 12.0 months, and the postoperative 0.5-, 1-, and 3-year DSS rates were 71.4%, 46.4% and 9.8%, respectively. We applied both the Cox proportional hazards model and the decision tree model to determine the prognostic factors of primary GSM. The Cox proportional hazards model demonstrated that age at presentation, tumour size, metastasis state and adjuvant chemotherapy (CT) were independent prognostic factors for DSS. The decision tree model suggested that age <71 years and adjuvant CT were associated with a better prognosis for GSM patients. The nomogram generated via the Cox proportional hazards model was developed by applying the rms package in R version 3.5.0. The C-index of internal validation for DSS prediction was 0.67 (95% confidence interval (CI), 0.63 to 0.70). The calibration curve at one year suggested that there was good consistency between the predicted DSS and the actual DSS probability. This study was the first to develop a disease-specific nomogram for predicting the prognosis of primary GSM patients after craniotomy, which can help clinicians immediately and accurately predict patient prognosis and conduct further treatment.

Abstract 5033: The role of the epigenetic regulator SATB2 in colon cancer progression

Abstract SATB2 is a nuclear matrix-associated transcription factor that orchestrates gene expression by regulating higher-order chromatin structure. Using antibody-based screening of 48 normal human tissues and 20 cancer types, we identified SATB2 as a protein almost exclusively expressed in gastrointestinal tissue. We confirmed this observation at the mRNA level in 10,533 human tissue samples. Differential expression of SATB2 was observed in colorectal cancer, with loss of expression occurring along the adenoma-carcinoma sequence. Additionally, SATB2 expression was markedly decreased in metastatic SW620 colon cancer cells and siRNA knockdown of SATB2 expression in parental SW480 cells increased their growth and migratory capacity. Ectopic expression of SATB2 in the metastatic variant reversed the observed phenotype. Using tissue microarray and automated image analysis of SATB2 expression in colorectal cancers (n=309), SATB2 was demonstrated by multivariate Cox regression analysis to be an independent predictor of disease-specific survival (HR=0.52, 95% CI 0.32-0.83, p=0.006). SATB2 mRNA levels were examined in a second cohort (n=290) and again, SATB2 was demonstrated to be an independent predictor of disease-specific survival (HR=0.40, 95% CI 0.18-0.92, p=0.031). Interestingly, in colorectal cancer patients, SATB2 levels significantly correlated with CD3+ T-cell infiltrates in the tumours (p=0.006) and inversely correlated with COX2 expression (p=0.019). In two independent colorectal cancer cohorts (n=467), SATB2-low tumours were found to be significantly enriched in the CMS1 (immune-related) subtype (p&amp;lt;0.001). Gene set enrichment analysis revealed that SATB2 low tumours demonstrated altered immune signalling with significant increases in IFNγ (p=0.001), IL6 (p=0.001), IL8 (p&amp;lt;0.001) TFGβ (p&amp;lt;0.001), which was mirrored by in vitro manipulation of SATB2 expression in colon cancer cells. Furthermore, in a longitudinal cohort of patients with ulcerative colitis, we observed a significant correlation between loss of SATB2 expression and occurrence of future cancers (p=0.013). We therefore postulate that SATB2 acts as a master regulator of the inflammatory response in the gut and loss of expression is significantly associated with the progression of colorectal cancer. Citation Format: Donal J. Brennan, Kirsha Naicker, Sudipto Das, Bruce Moran, Rut Klinger, Fredrik Ponten, Stephen Hewitt, Karin Jirström, Annette T. Byrne, Jacintha O'Sullivan, William M. Gallagher, Darran P. O'Connor. The role of the epigenetic regulator SATB2 in colon cancer progression [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 5033.