Simon Turcotte

Tumor MHC Class I Expression Improves the Prognostic Value of T-cell Density in Resected Colorectal Liver Metastases

Using a tissue microarray of liver metastases from 158 patients with colorectal cancer, Turcotte and colleagues show that high MHC class I expression with dense intratumoral T-cell infiltration identifies patients with favorable outcomes independent of conventional prognostic factors. Tumor-infiltrating lymphocytes (TIL) in colorectal cancer liver metastases (CLM) have been associated with more favorable patient outcomes, but whether MHC class I (MHC-I) expression on cancer cells affects prognosis is uncertain. Immunohistochemistry was performed on a tissue microarray of 158 patients with CLM, who underwent partial hepatectomy with curative intent. Using the antibody HC-10, which detects HLA-B and HLA-C antigens and a minority of HLA-A antigens, MHC-I expression was correlated with β-2 microglobulin (β2m; r = 0.7; P < 0.001), but not with T-cell density (r < 0.32). The median follow-up for survivors was 9.7 years. High levels of MHC-I expression in tumors concomitant with high T-cell infiltration (CD3, CD4, or CD8) best identified patients with favorable outcomes, compared with patients with one or none of these immune features. The median overall survival (OS) of patients with MHC-IhiCD3hi tumors (n = 31) was 116 months compared with 40 months for the others (P = 0.001), and the median time to recurrence (TTR) was not reached compared with 17 months (P = 0.008). By multivariate analysis, MHChiCD3hi was associated with OS and TTR independent of the standard clinicopathologic variables. An immune score that combines MHC-I expression and TIL density may be a valuable prognostic tool in the treatment of patients with CLM. Cancer Immunol Res; 2(6); 530–7. ©2014 AACR.

Is intra-operative ultrasound still useful for the detection of a hepatic tumour in the era of modern pre-operative imaging?

BACKGROUND The current role of intra-operative ultrasound (IOUS) is questioned because of recent progress in medical imaging. The aim of the present study was to determine the accuracy of IOUS in the detection of a hepatic tumour (HT) compared with a pre-operative multi-detector computed tomography (MDCT) scan and magnetic resonance imaging (MRI). METHODS This retrospective study included 418 patients evaluated using an 8-slice MDCT scan (SCAN8), 64-slice MDCT scan (SCAN64) and MRI alone or combined with a computed tomography (CT) scan. The pathological result was used as a gold standard. RESULTS Correlation rates for the number of detected lesions compared with pathology results were 0.627 for SCAN8, 0.785 for SCAN64, 0.657 for MRI and 0.913 for IOUS. Compared with pathology, the rate of concordance was significantly higher with IOUS (0.871) than with SCAN8 (0.736; P=0.011), SCAN64 (0.792; P<0.001) and MRI (0.742; P<0.001). IOUS was responsible for a change in operative strategy in 16.5% of patients. Surgery was extended in 12.4%, limited in 1.7% and abandoned in 2.4%. CONCLUSIONS Compared with cross-sectional pre-operative imaging, IOUS is still superior for the detection of HT and the planning of surgery. IOUS remains recommended as a routine procedure in patients having a hepatic resection in the era of modern pre-operative imaging.

Deep Learning: A Primer for Radiologists

Deep learning is a class of machine learning methods that are gaining success and attracting interest in many domains, including computer vision, speech recognition, natural language processing, and playing games. Deep learning methods produce a mapping from raw inputs to desired outputs (eg, image classes). Unlike traditional machine learning methods, which require hand-engineered feature extraction from inputs, deep learning methods learn these features directly from data. With the advent of large datasets and increased computing power, these methods can produce models with exceptional performance. These models are multilayer artificial neural networks, loosely inspired by biologic neural systems. Weighted connections between nodes (neurons) in the network are iteratively adjusted based on example pairs of inputs and target outputs by back-propagating a corrective error signal through the network. For computer vision tasks, convolutional neural networks (CNNs) have proven to be effective. Recently, several clinical applications of CNNs have been proposed and studied in radiology for classification, detection, and segmentation tasks. This article reviews the key concepts of deep learning for clinical radiologists, discusses technical requirements, describes emerging applications in clinical radiology, and outlines limitations and future directions in this field. Radiologists should become familiar with the principles and potential applications of deep learning in medical imaging. ©RSNA, 2017.

Resection versus transplantation for early hepatocellular carcinoma: more art than science.

H epatocellular carcinoma (HCC) is the second (in men) and sixth (in women) most frequent cause of cancer death worldwide.1 In the United Sates, the incidence of this malignancy has tripled since 1975, and approximately 26,000 new patients will be diagnosed in 2012.2 Transplantation has become a major weapon in the arsenal against HCC. There is considerable rationale for liver transplantation as the primary therapy for early-stage HCC. First, removal of the entire liver eliminates occult intrahepatic metastatic disease and the risk of a second primary cancer developing in a chronically diseased liver. Although there have been major technical advances in partial hepatectomy, it generally cannot be performed safely in patients with advanced cirrhosis because of associated portal hypertension. Since the late 1990s, numerous centers worldwide have reported that overall survival after liver transplantation is just slightly inferior for patients with early-stage HCC (within Milan Criteria—1 tumor ≤5 cm or up to 3 tumors ≤3 cm and no vascular invasion) than for those who underwent transplantation for benign disease.3 The approximate benchmark is 70% overall survival at 5 years.4 Of course, there are several challenges for liver transplantation in HCC. One of the major impediments to the broad application of liver transplantation is the shortage of donor organs. Consequently, there is often a substantial dropout rate on the waiting list due to progression of either the cancer or the underlying cirrhosis. Other issues are the high cost of transplantation and the chronic immunosuppression required postoperatively. Given the limited availability of donor organs and poor outcomes of patients dropping off the waiting list, judicious organ allocation has become paramount. In many countries, liver transplantation is favored for patients with early HCC and poor liver function, whereas liver resection is performed for patients with preserved hepatic function. A 5-year survival of 50% to 70% has been reported in series in cirrhotic patients treated with resection for HCC within Milan Criteria.5–8 In this issue of Annals of Surgery, Adam and colleagues performed a retrospective analysis of patients with a solitary HCC of size 5 cm or less.9 They treated 198 patients with partial hepatectomy or transplantation in a nonrandomized fashion at Hôpital Paul Brousse between 1990 and 2010. Similar to most retrospective reviews, patient selection could not be precisely defined, which somewhat limits the interpretation of the results. Therapy was based on what was felt to “be the best curative strategy for an individual patient” and resection tended to be performed for peripheral tumors in Child-Pugh class A patients. Partial hepatectomy was offered to 98 patients and performed in 97 (1% dropout rate) and liver transplantation planned for 111 and occurred in 101 patients (9% dropout rate). As would be expected, the clinicopathologic features of the 2 groups were quite different (Table 1). Most (88%) patients treated with partial hepatectomy were Child-Pugh class A with larger, poorly differentiated tumors, whereas patients treated with transplantation were Child-Pugh B (45%) or C (37%) and younger and had more tumors on final pathologic examination. For unclear reasons, 81% of the patients in the partial hepatectomy group underwent preoperative hepatic artery embolization. Overall, the average delay from diagnosis to liver resection was 7 months ± 7 months. In the entire population, the 3and 5-year overall survival was 68% and 51% for partial hepatectomy and 74% and 70% for transplantation, respectively (Figure 3, P = 0.01), using intentionto-treat methodology (ie, including dropouts). Recurrence-free survival was also worse with partial hepatectomy. The authors have certainly achieved excellent results with minimal perioperative mortality. Their survival and recurrence outcomes with partial hepatectomy or transplantation are consistent with data reported in many published series.2,3 However, the authors ventured further and compared

A Validated Prognostic Multigene Expression Assay for Overall Survival in Resected Colorectal Cancer Liver Metastases.

Purpose: Risk stratification after surgery for colorectal cancer liver metastases (CRLM) is achieved using clinicopathologic variables, however, is of limited accuracy. We sought to derive and externally validate a multigene expression assay prognostic of overall survival (OS) that is superior to clinicopathologic variables in patients with surgically resected CRLM. Experimental Design: We measured mRNA expression in prospectively collected frozen tumor from 96 patients with surgically resected CRLM at Memorial Sloan Kettering Cancer Center (MSKCC, New York, NY). We retrospectively generated a 20-gene molecular risk score (MRS) and compared its prognostic utility for OS and recurrence-free survival (RFS) with three common clinical risk scores (CRS). We then tested the prognostic ability of the MRS in an external validation cohort (European) of 119 patients with surgically resected CRLM at the University Medical Center Utrecht (Utrecht, the Netherlands) and Paul Brousse Hospital (Villejuif, France). Results: For OS in the MSKCC cohort, MRS was the strongest independent prognosticator (HR, 3.7–4.9; P < 0.001) followed by adjuvant chemotherapy (HR, 0.3; P ≤ 0.001). For OS in the European cohort, MRS was the only independent prognosticator (HR, 3.5; P = 0.007). For RFS, MRS was also independently prognostic in the MSKCC cohort (HR, 2.4–2.6; P ≤ 0.001) and the European cohort (HR, 1.6–2.5; P ≤ 0.05). Conclusions: Compared with CRSs, the MRS is more accurate, broadly applicable, and an independent prognostic biomarker of OS in resected CRLM. This MRS is the first externally validated prognostic multigene expression assay after metastasectomy for CRLM and warrants prospective validation. Clin Cancer Res; 22(10); 2575–82. ©2016 AACR.

Pharmacokinetics of intraperitoneal irinotecan in a pig model

Complete surgical cytoreduction of peritoneal implants and immediate intraperitoneal (lP) chemotherapy offers the greatest survival in selected patients with peritoneal carcinomatosis. This study was undertaken to describe the metabolism and pharmacokinetics of normothermic intraperitoneal CPT‐11, free and glucuronized SN‐38, in a pig model. Thirteen pigs were used for experimentation. Animals were grouped for IV and IP CPT‐1 1 administration. Eleven pigs underwent laparotomy through a midline incision and instillation of 100, 200, and 400mg IP CPT‐11. Systemic venous blood, portal blood and peritoneal fluid samples were taken at 5, 10, 20, 30, and 45 min, then every hour up to 8 hr for the 100 mg. For the three groups, peritoneal CPT‐11 exposition was on average 4.9 times greater in the peritoneum than in the systemic venous or portal circulations and the systemic CPT‐11 fraction absorbed from the peritoneum linearly increased with time. Free SN‐38 was measurable in the earliest peritoneal samples taken. The initial instillation dose of CPT‐11 did not impact on the SN‐38 converted fraction, which remained stable at approximately 0.04% during the first 4 hour. Mean peritoneal SN‐38: CPT‐11 AUC ratio was 0.043. OPT‐11 peritoneal conversion into SN‐38 appeared slightly Inferior to the systemic conversion ratio. This norrnothermic IP OPT‐11 pharmacokinetic study performed in a pig model confirms the possibility to achieve at least a 30 times higher peritoneal than systemic exposure. Peritoneal exposure to active SN‐38 begins at the moment of CPT‐11 peritoneal instillation. A fixed and small traction of less than 0.1% of CPT‐11 is converted into SN‐38, underlying the importance of a sufficient initial IP dose of CPT‐11. J. Surg. Oncol. 2010; 101:637–642.

LI-RADS for MR Imaging Diagnosis of Hepatocellular Carcinoma: Performance of Major and Ancillary Features

Purpose To evaluate the performance of major features, ancillary features, and categories of Liver Imaging Reporting and Data System (LI-RADS) version 2014 at magnetic resonance (MR) imaging for the diagnosis of hepatocellular carcinoma (HCC). Materials and Methods This retrospective institutional review board-approved study included patients with liver MR imaging and at least one pathologically proved lesion. Between 2004 and 2016, 102 patients (275 observations including 113 HCCs) met inclusion criteria. Two radiologists independently assessed major and ancillary imaging features for each liver observation and assigned a LI-RADS category. Per-lesion estimates of diagnostic performance of major features, ancillary features, and LI-RADS categories were assessed by using generalized estimating equation models. Results Major features (arterial phase hyperenhancement, washout, capsule, and threshold growth) had a sensitivity of 88.5%, 60.6%, 32.9%, and 41.6%, and a specificity of 18.6%, 84.8%, 98.8%, and 83.2% for HCC, respectively. Ancillary features (mild-moderate T2 hyperintensity, restricted diffusion, mosaic architecture, intralesional fat, lesional fat sparing, blood products, and subthreshold growth) had a sensitivity of 62.2%, 54.8%, 9.9%, 30.9%, 23.1%, 2.8%, and 48.3%, and a specificity of 79.4%, 90.6%, 99.4%, 94.2%, 83.1%, 99.3%, and 91.4% for HCC, respectively. The LR-5 or LR-5 V categories had a per-lesion sensitivity of 50.8% and a specificity of 95.8% for HCC, respectively. The LR-4, LR-5, or LR-5 V categories (determined by using major features only vs combination of major and ancillary features) had a per-lesion sensitivity of 75.9% and 87.9% and a per-lesion specificity of 87.5% and 86.2%, respectively. Conclusion The use of ancillary features in combination with major features increases the sensitivity while preserving a high specificity for the diagnosis of HCC.

Prediction of Hepatocellular Carcinoma Recurrence Beyond Milan Criteria After Resection: Validation of a Clinical Risk Score in an International Cohort

Objective: This study aims to validate a previously reported recurrence clinical risk score (CRS). Summary of Background Data: Salvage transplantation after hepatocellular carcinoma (HCC) resection is limited to patients who recur within Milan criteria (MC). Predicting recurrence patterns may guide treatment recommendations. Methods: An international, multicenter cohort of R0 resected HCC patients were categorized by MC status at presentation. CRS was calculated by assigning 1 point each for initial disease beyond MC, multinodularity, and microvascular invasion. Recurrence incidences were estimated using competing risks methodology, and conditional recurrence probabilities were estimated using the Bayes theorem. Results: From 1992 to 2015, 1023 patients were identified, of whom 613 (60%) recurred at a median follow-up of 50 months. CRS was well validated in that all 3 factors remained independent predictors of recurrence beyond MC (hazard ratio 1.5–2.1, all P < 0.001) and accurately stratified recurrence risk beyond MC, ranging from 19% (CRS 0) to 67% (CRS 3) at 5 years. Among patients with CRS 0, no other factors were significantly associated with recurrence beyond MC. The majority recurred within 2 years. After 2 years of recurrence-free survival, the cumulative risk of recurrence beyond MC within the next 5 years for all patients was 14%. This risk was 12% for patients with initial disease within MC and 17% for patients with initial disease beyond MC. Conclusions: CRS accurately predicted HCC recurrence beyond MC in this international validation. Although the risk of recurrence beyond MC decreased over time, it never reached zero.

Benign gastrobronchial fistula with adenocarcinoma of the right mainstem bronchus

Fistulization of the gastric conduit into the tracheobronchial tree is a potential complication after esophageal reconstruction. In this situation, the effects of the refluxate on the mucosa of the tracheobronchial tree is unknown. This article presents the 11-year evolution of a patient who tolerated a fistulous communication between the gastric interposition and the right mainstem bronchus, resulting in an adenocarcinoma of the bronchus.

Abstract 4670: Evidence of neoantigen-reactive T cell response in a case of relapsing, mismatch-repair gene proficient, colorectal cancer

Whether the endogenous T cell reactivity to antigens derived from cancer mutations, called neoantigens (neoAgs), can be exploited for immunotherapy in patients with mismatch-repair (MMR) gene proficient metastatic colorectal cancer (CRC) is largely unknown. Here, we investigated the frequency of neoAg-reactive T cells longitudinally in a patient with relapsing CRC. Using whole exome and RNA sequencing, we identified 96 single nucleotide variant (SNV), 8 frameshift and 29 insertion/deletion mutations in a liver metastasis. Co-culture assays between in vitro expanded tumor infiltrating T lymphocytes (TILs) and autologous CD40-activated B cells loaded with peptides derived from SNVs led to the identification of two CD8+ T cell clones specific for PAMA733E, and respectively 1, 4 and 2 CD4+ T cell clones specific for PABPC1G563S, PDE4DIPR685S and TRPM4A480V. T cell clonality was confirmed by TCR Sanger sequencing. We assessed the in vivo frequency of these neoAg-reactive T cell clones by TCRβ chain deep sequencing (Adaptive Biotechnologies). The dominant T cell clone, reactive to PAMA733E, represented 5.2% of TILs in the primary tumor resected in 2010, 2.2% of TILs in the colonic recurrence resected in 2011, and 3.2% of TILs in the liver metastasis resected in 2012. The neoAg-reactive T cells were also detected in the peritumoral liver, but not into the distant normal liver. Along the disease course, most reactive T cell clones were not detected in the tumor draining lymph nodes or in the peripheral blood, at an average detection capacity of 1/3680 and 1/185,495 T cells respectively. When compared to the RNAseq data from 60 other CRC liver metastases, the relatively high level of transcripts related to immune cells, antigen processing and presentation, IFN-γ responsive genes, T-cell inhibitory and stimulatory receptors, cytokines, and chemokines observed in our patient metastasis suggested that there was an ongoing spontaneous immune response intratumoraly, co-existing with many immune-suppressive molecules. Our results support that neoAg-reactive T cells can be found in non-highly mutated, MMR proficient CRC tumors, at a much higher frequency than in the peripheral blood or the draining lymph nodes. As an adjuvant strategy to prevent recurrence, it may be possible in some patients to boost the immune response against a relevant neoAg expressed in the primary tumor. At the metastatic stage, a broader array of neoAgs may be targetable. Citation Format: Melissa Mathieu, Alexandre Paradis, Sandy Pelletier, Steven Hebert, Kevin Boutin, Eric Audemard, Sylvie Mader, Claudia Kleinman, Simon Turcotte. Evidence of neoantigen-reactive T cell response in a case of relapsing, mismatch-repair gene proficient, colorectal cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 4670.