Faculty Bibliography

Patients often respond differently to a treatment because of individual heterogeneity. Failures of clinical trials can be substantially reduced if, prior to an investigational treatment, patients are stratified into responders and nonresponders based on biological or demographic characteristics. These characteristics are captured by a predictive signature. In this paper, we propose a procedure to search for predictive signatures based on the approach of patient rule induction method. Specifically, we discuss selection of a proper objective function for the search, present its algorithm, and describe a resampling scheme that can enhance search performance. Through simulations, we characterize conditions under which the procedure works well. To demonstrate practical uses of the procedure, we apply it to two real-world data sets. We also compare the results with those obtained from a recent regression-based approach, Adaptive Index Models, and discuss their respective advantages. In this study, we focus on oncology applications with survival responses.

BACKGROUND: Identification of primary melanoma patients at the highest risk of recurrence remains a critical challenge, and monitoring for recurrent disease is limited to costly imaging studies. We recently reported our array-based discovery of prognostic serum miRNAs in melanoma. In the current study, we examined the clinical utility of these serum-based miRNAs for prognosis as well as detection of melanoma recurrence. METHODS: Serum levels of 12 miRNAs were tested using qRT-PCR at diagnosis in 283 melanoma patients (training cohort, n = 201; independent validation, n = 82; median follow-up, 68.8 months). A refined miRNA signature was chosen and evaluated. We also tested the potential clinical utility of the miRNAs in early detection and monitoring of recurrence using multiple longitudinal samples (pre- and postrecurrence) in a subset of 82 patients (n = 225). In addition, we integrated our miRNA signature with publicly available Cancer Genome Atlas data to examine the relevance of these miRNAs to melanoma biology. RESULTS: Four miRNAs (miR-150, miR-30d, miR-15b, and miR-425) in combination with stage separated patients by recurrence-free survival (RFS) and overall survival (OS) and improved prediction of recurrence over stage alone in both the training and validation cohorts (training RFS and OS, P < .001; validation RFS, P < .001; OS, P = .005). Serum miR-15b levels significantly increased over time in recurrent patients (P < .001), adjusting for endogenous controls as well as age, sex, and initial stage. In nonrecurrent patients, miR-15b levels were not significantly changed with time (P =.17). CONCLUSIONS: Data demonstrate that serum miRNAs can improve melanoma patient stratification over stage and support further testing of miR-15b to guide patient surveillance. Cancer 2015;121:51-59. (c) 2014 American Cancer Society.

Current surgical treatment of primary melanoma is uniform for all histosubtypes, although certain types of melanoma, such as acral lentiginous melanoma (ALM), have a worse prognosis. No study has explored the effectiveness of standard melanoma treatment guidelines for managing ALM compared with nonacral melanoma (NAM). Study subjects were identified from a prospectively enrolled database of patients with primary melanoma at New York University. Patients with ALM were matched to those with NAM (1:3) by gender and melanoma stage, including substage (ALM, 61; NAM, 183). All patients received standard-of-care treatment. Recurrence and survival outcomes in both cohorts were compared. ALM histologic subtype was an independent negative predictor of recurrence-free survival (hazard ratio [HR], 2.24; P=.001) and melanoma-specific survival (HR, 2.58; P=.001) compared with NAM. Recurrence was significantly more common in patients with ALM than in those with NAM (49% vs 30%; P=.007). For tumors less than 2 mm in thickness, a significantly higher recurrence rate was seen with ALM versus NAM (P=.048). No significant difference was seen in recurrence for tumors greater than 2 mm (P=.12). Notably, the rate of locoregional recurrence was nearly double for ALM compared with NAM (P=.001). The data presented herein reveal a high rate of locoregional failure in ALM compared with NAM when controlling for AJCC stage. These results raise the question of whether ALM may require more aggressive surgical treatment than nonacral cutaneous melanomas of equal thickness, particularly in tumors less than 2 mm thick. Larger multicenter trials are necessary for further conclusions.

BACKGROUND:Biomarkers predicting tumor response are important to emerging targeted therapeutics. Complimentary methods to assess and understand genetic changes and heterogeneity within only few cancer cells in tissue will be a valuable addition for assessment of tumors such as prostate cancer that often have insufficient tumor for next generation sequencing in a single biopsy core. METHODS:Using confocal microscopy to identify cell-to-cell relationships in situ, we studied the most common gene rearrangement in prostate cancer (TMPRSS2 and ERG) and the tumor suppressor CHD1 in 56 patients who underwent radical prostatectomy. RESULTS:Wild type ERG was found in 22 of 56 patients; ERG copy number was increased in 10/56, and ERG rearrangements confirmed in 24/56 patients. In 24 patients with ERG rearrangements, the mechanisms of rearrangement were heterogeneous, with deletion in 14/24, a split event in 7/24, and both deletions and split events in the same tumor focus in 3/24 patients. Overall, 14/45 (31.1%) of patients had CHD1 deletion, with the majority of patients with CHD1 deletions (13/14) correlating with ERG-rearrangement negative status (P < 0.001). CONCLUSIONS:These results demonstrate the ability of confocal microscopy and FISH to identify the cell-to-cell differences in common gene fusions such as TMPRSS2-ERG that may arise independently within the same tumor focus. These data support the need to study complimentary approaches to assess genetic changes that may stratify therapy based on predicted sensitivities.

Neural infiltration in primary melanoma is a histopathologic feature that has been associated with desmoplastic histopathologic subtype and local recurrence in the literature. We tested the hypothesis that improved detection and characterization of neural infiltration into peritumoral or intratumoral location and perineural or intraneural involvement could have a prognostic relevance. We studied 128 primary melanoma cases prospectively accrued and followed at New York University using immunohistochemical detection with antihuman neurofilament protein and routine histology with hematoxylin and eosin. Neural infiltration, defined as the presence of tumor cells involving or immediately surrounding nerve foci, was identified and characterized using both detection methods. Neural infiltration rate of detection was enhanced by immunohistochemistry for neurofilament in matched-pair design (47% by immunohistochemistry versus 25% by routine histology). Immunohistochemical detection of neural infiltration was significantly associated with ulceration (P = .021), desmoplastic and acral lentiginous histologic subtype (P = .008), and head/neck/hands/feet tumor location (P = .037). Routinely detected neural infiltration was significantly associated with local recurrence (P = .010). Immunohistochemistry detected more intratumoral neural infiltration cases compared with routine histology (30% versus 3%, respectively). Peritumoral and intratumoral nerve location had no impact on clinical outcomes. Using a multivariate model controlling for stage, neither routinely detected neural infiltration nor enhanced immunohistochemical characterization of neural infiltration was significantly associated with disease-free or overall survival. Our data demonstrate that routinely detected neural infiltration is associated with local recurrence in all histologic subtypes but that improved detection and characterization of neural infiltration with immunohistochemistry in primary melanoma does not add to prognostic relevance.

Previous reports have demonstrated a role for hedgehog signaling in melanoma progression, prompting us to explore the therapeutic benefit of targeting this pathway in melanoma. We profiled a panel of human melanoma cell lines and control melanocytes for altered expression of hedgehog pathway members and determined the consequences of both genetic and pharmacological inhibition of the hedgehog pathway activator Smoothened (SMO) in melanoma, both in vitro and in vivo. We also examined the relationship between altered expression of hedgehog pathway mediators and survival in a well-characterized cohort of metastatic melanoma patients with prospectively collected follow up information. Studies revealed that over 40% of the melanoma cell lines examined harbored significantly elevated levels of the hedgehog pathway mediators SMO, GLI2, and PTCH1 compared to melanocytes (p < 0.05). SMO inhibition using siRNA and the small molecule inhibitor, NVP-LDE-225, suppressed melanoma growth in vitro, particularly in those cell lines with moderate SMO and GLI2 expression. NVP-LDE-225 also induced apoptosis in vitro and inhibited melanoma growth in a xenograft model. Gene expression data also revealed evidence of compensatory up-regulation of two other developmental pathways, Notch and WNT, in response to hedgehog pathway inhibition. Pharmacological and genetic SMO inhibition also downregulated genes involved in human embryonic stem cell pluripotency. Finally, increased SMO expression and decreased expression of the hedgehog pathway repressor GLI3 correlated with shorter post recurrence survival in metastatic melanoma patients. Our data demonstrate that hedgehog pathway inhibition might be a promising targeted therapy in appropriately selected metastatic melanoma patients.