NAS-SGAN: A Semi-Supervised Generative Adversarial Network Model for Atypia Scoring of Breast Cancer Histopathological Images.

Nuclear atypia scoring (NAS), forms a significant factor in determining individualized treatment plans and also for the prognosis of the disease. Automation of cancer grading using quantitative image-based analysis of histopathological images can circumvent the shortcomings of the prevailing manual grading and can assist the pathologists in cancer diagnosis. However, developing such a robust classifier model require sufficient amount of annotated data, while the labeled histopathological images are scarce and expensive to procure as annotation forms a time-consuming and laborious task. Hence, a semi-supervised learning framework combined with the deep neural network based generative adversarial training, that can improve the performance of the classification model with limited annotated data, is proposed in this paper. The proposed NAS-SGAN model consists of discriminator and generator models that are trained in an adversarial manner using both labeled and unlabeled samples. The discriminator model is designed as an unsupervised model stacked over the supervised model sharing the model parameters and learns the data distribution by extracting the discriminative features. The generator model is trained over a stable feature matching objective function following a composite GAN architecture, and its for the first time the semi-supervised GAN model is explored for the grading of breast cancer. Experimental analysis shows that the proposed model could better discriminate different cancer grades thereby improving the robustness and accuracy of the system, even with limited amount of labeled samples.

Batch Mode Active Learning on the Riemannian Manifold for Automated Scoring of Nuclear Pleomorphism in Breast Cancer.

Breast cancer is the most prevalent invasive type of cancer among women. The mortality rate of the disease can be reduced considerably through timely prognosis and felicitous treatment planning, by utilizing the computer aided detection and diagnosis techniques. With the advent of whole slide image (WSI) scanners for digitizing the histopathological tissue samples, there is a drastic increase in the availability of digital histopathological images. However, these samples are often unlabeled and hence they need labeling to be done through manual annotations by domain experts and experienced pathologists. But this annotation process required for acquiring high quality large labeled training set for nuclear atypia scoring is a tedious, expensive and time consuming job. Active learning techniques have achieved widespread acceptance in reducing this human effort in annotating the data samples. In this paper, we explore the possibilities of active learning on nuclear pleomorphism scoring over a non-Euclidean framework, the Riemannian manifold. Active learning technique adopted for the cancer grading is in the batch-mode framework, that adaptively identifies the apt batch size along with the batch of instances to be queried, following a submodular optimization framework. Samples for annotation are selected considering the diversity and redundancy between the pair of samples, based on the kernelized Riemannian distance measures such as log-Euclidean metrics and the two Bregman divergences - Stein and Jeffrey divergences. Results of the adaptive Batch Mode Active Learning on the Riemannian metric show a superior performance when compared with the state-of-the-art techniques for breast cancer nuclear pleomorphism scoring, as it makes use of the information from the unlabeled samples.

Computer-Aided Histopathological Image Analysis Techniques for Automated Nuclear Atypia Scoring of Breast Cancer: a Review.

Breast cancer is the most common type of malignancy diagnosed in women. Through early detection and diagnosis, there is a great chance of recovery and thereby reduce the mortality rate. Many preliminary tests like non-invasive radiological diagnosis using ultrasound, mammography, and MRI are widely used for the diagnosis of breast cancer. However, histopathological analysis of breast biopsy specimen is inevitable and is considered to be the golden standard for the affirmation of cancer. With the advancements in the digital computing capabilities, memory capacity, and imaging modalities, the development of computer-aided powerful analytical techniques for histopathological data has increased dramatically. These automated techniques help to alleviate the laborious work of the pathologist and to improve the reproducibility and reliability of the interpretation. This paper reviews and summarizes digital image computational algorithms applied on histopathological breast cancer images for nuclear atypia scoring and explores the future possibilities. The algorithms for nuclear pleomorphism scoring of breast cancer can be widely grouped into two categories: handcrafted feature-based and learned feature-based. Handcrafted feature-based algorithms mainly include the computational steps like pre-processing the images, segmenting the nuclei, extracting unique features, feature selection, and machine learning-based classification. However, most of the recent algorithms are based on learned features, that extract high-level abstractions directly from the histopathological images utilizing deep learning techniques. In this paper, we discuss the various algorithms applied for the nuclear pleomorphism scoring of breast cancer, discourse the challenges to be dealt with, and outline the importance of benchmark datasets. A comparative analysis of some prominent works on breast cancer nuclear atypia scoring is done using a benchmark dataset which enables to quantitatively measure and compare the different features and algorithms used for breast cancer grading. Results show that improvements are still required, to have an automated cancer grading system suitable for clinical applications.

Sparse Representation Over Learned Dictionaries on the Riemannian Manifold for Automated Grading of Nuclear Pleomorphism in Breast Cancer.

Breast cancer is found to be the most pervasive type of cancer among women. Computer aided detection and diagnosis of cancer at the initial stages can increase the chances of recovery and thus reduce the mortality rate through timely prognosis and adequate treatment planning. The nuclear atypia scoring or histopathological breast tumor grading remains to be a challenging problem due to the various artifacts and variabilities introduced during slide preparation and also because of the complexity in the structure of the underlying tissue patterns. Inspired by the success of symmetric positive definite (SPD) matrices in many of the challenging tasks in machine learning and computer vision, a sparse coding and dictionary learning on SPD matrices is proposed in this paper for the breast tumor grading. The proposed covariance-based SPD matrices form a Riemannian manifold and are represented as the sparse combination of Riemannian dictionary atoms. Non-linearity of the SPD manifold is tackled by embedding into the reproducing kernel Hilbert space using kernels derived from log-Euclidean metric, Jeffrey and Stein divergences and compared with the non-kernel-based affine invariant Riemannian metric. The novelty of the work lies in exploiting the kernel approach for the Hilbert space embedding of the Riemannian manifold, that can achieve a better discrimination of the breast cancer tissues, following a sparse representation over learned dictionaries and henceforth it outperforms many of the state-of-the-art algorithms in breast cancer grading in terms of quantitative and qualitative analysis.

Molecular Analyses Support the Safety and Activity of Retroviral Replicating Vector Toca 511 in Patients.

Purpose: Toca 511 is a gammaretroviral replicating vector encoding cytosine deaminase that selectively infects tumor cells and converts the antifungal drug 5-fluorocytosine into the antineoplastic drug 5-fluorouracil, which directly kills tumor cells and stimulates antitumor immune responses. As part of clinical monitoring of phase I clinical trials in recurrent high-grade glioma, we have performed extensive molecular analyses of patient specimens to track vector fate.Patients and Methods: Toca 511 and Toca FC (extended-release 5-fluorocytosine) have been administered to 127 high-grade glioma patients across three phase I studies. We measured Toca 511 RNA and DNA levels in available body fluids and tumor samples from patients to assess tumor specificity. We mapped Toca 511 integration sites and sequenced integrated Toca 511 genomes from patient samples with detectable virus. We measured Toca 511 levels in a diverse set of tissue samples from one patient.Results: Integrated Toca 511 is commonly detected in tumor samples and is only transiently detected in blood in a small fraction of patients. There was no believable evidence for clonal expansion of cells with integrated Toca 511 DNA, or preferential retrieval of integration sites near oncogenes. Toca 511 sequence profiles suggest most mutations are caused by APOBEC cytidine deaminases acting during reverse transcription. Tissue samples from a single whole-body autopsy affirm Toca 511 tumor selectivity.Conclusions: Toca 511 and Toca FC treatment was not associated with inappropriate integration sites and clonal expansion. The vector is tumor-selective and persistent in patients who received Toca 511 injections. Clin Cancer Res; 24(19); 4680-93. ©2018 AACR.

Durable complete responses in some recurrent high-grade glioma patients treated with Toca 511 + Toca FC.

Background: Vocimagene amiretrorepvec (Toca 511) is an investigational gamma-retroviral replicating vector encoding cytosine deaminase that, when used in combination with extended-release 5-fluorocytosine (Toca FC), results preclinically in local production of 5-fluorouracil, depletion of immune-suppressive myeloid cells, and subsequent induction of antitumor immunity. Recurrent high-grade glioma (rHGG) patients have a high unmet need for effective therapies that produce durable responses lasting more than 6 months. In this setting, relapse is nearly universal and most responses are transient. Methods: In this Toca 511 ascending-dose phase I trial (NCT01470794), HGG patients who recurred after standard of care underwent surgical resection and received Toca 511 injected into the resection cavity wall, followed by orally administered cycles of Toca FC. Results: Among 56 patients, durable complete responses were observed. A subgroup was identified based on Toca 511 dose and entry requirements for the follow-up phase III study. In this subgroup, which included both isocitrate dehydrogenase 1 (IDH1) mutant and wild-type tumors, the durable response rate is 21.7%. Median duration of follow-up for responders is 35.7+ months. As of August 25, 2017, all responders remain in response and are alive 33.9+ to 52.2+ months after Toca 511 administration, suggesting a positive association of durable response with overall survival. Conclusions: Multiyear durable responses have been observed in rHGG patients treated with Toca 511 + Toca FC in a phase I trial, and the treatment will be further evaluated in a randomized phase III trial. Among IDH1 mutant patients treated at first recurrence, there may be an enrichment of complete responders.

Melanomas prevent endothelial cell death under restrictive culture conditions by signaling through AKT and p38 MAPK/ ERK-1/2 cascades.

Although melanoma progression and staging is clinically well characterized, a large variation is observed in pathogenesis, progression, and therapeutic responses. Clearly, intrinsic characteristics of melanoma cells contribute to this variety. An important factor, in both progression of the disease and response to therapy, is the tumor-associated vasculature. We postulate that melanoma cells communicate with endothelial cells (ECs) in order to establish a functional and supportive blood supply. We investigated the angiogenic potential of human melanoma cell lines by monitoring the survival of ECs upon exposure to melanoma conditioned medium (CM), under restrictive conditions. We observed long-term (up to 72 h) EC survival under hypoxic conditions upon treatment with all melanoma CMs. No such survival effect was observed with the CM of melanocytes. The CM of pancreatic and breast tumor cell lines did not show a long-term survival effect, suggesting that the survival factor is specific to melanoma cells. Furthermore, all size fractions (up to < 1 kDa) of the melanoma CM induced long-term survival of ECs. The survival effect observed by the < 1 kDa fraction excludes known pro-angiogenic factors. Heat inactivation and enzymatic digestion of the CM did not inactivate the survival factor. Global gene expression and pathway analysis suggest that this effect is mediated in part via the AKT and p38 MAPK/ ERK-1/2 signaling axis. Taken together, these data indicate the production of (a) survival factor/s (< 1 kDa) by melanoma cell lines, which enables long-term survival of ECs and promotes melanoma-induced angiogenesis.

Lipid phosphatase SHIP2 functions as oncogene in colorectal cancer by regulating PKB activation.

Colorectal cancer (CRC) is the second most common cause of cancer-related death, encouraging the search for novel therapeutic targets affecting tumor cell proliferation and migration. These cellular processes are under tight control of two opposing groups of enzymes; kinases and phosphatases. Aberrant activity of kinases is observed in many forms of cancer and as phosphatases counteract such "oncogenic" kinases, it is generally assumed that phosphatases function as tumor suppressors. However, emerging evidence suggests that the lipid phosphatase SH2-domain-containing 5 inositol phosphatase (SHIP2), encoded by the INPPL1 gene, may act as an oncogene. Just like the well-known tumor suppressor gene Phosphatase and Tensin Homolog (PTEN) it hydrolyses phosphatidylinositol (3,4,5) triphosphate (PI(3,4,5)P3). However, unlike PTEN, the reaction product is PI(3,4)P2, which is required for full activation of the downstream protein kinase B (PKB/Akt), suggesting that SHIP2, in contrast to PTEN, could have a tumor initiating role through PKB activation. In this work, we investigated the role of SHIP2 in colorectal cancer. We found that SHIP2 and INPPL1 expression is increased in colorectal cancer tissue in comparison to adjacent normal tissue, and this is correlated with decreased patient survival. Moreover, SHIP2 is more active in colorectal cancer tissue, suggesting that SHIP2 can induce oncogenesis in colonic epithelial cells. Furthermore, in vitro experiments performed on colorectal cancer cell lines shows an oncogenic role for SHIP2, by enhancing chemoresistance, cell migration, and cell invasion. Together, these data indicate that SHIP2 expression contributes to the malignant potential of colorectal cancer, providing a possible target in the fight against this devastating disease.

Tissue inhibitor of metalloproteinase-3 (TIMP3) expression decreases during melanoma progression and inhibits melanoma cell migration.

AIMS: Malignant melanoma is the most aggressive form of skin cancer, and metastatic dissemination to regional and visceral sites is responsible for the majority of melanoma-related mortalities. In a recent study by our group, we observed reduced expression of tissue inhibitor of metalloproteinase-3 (TIMP3) in the majority of stage III melanoma samples studied. TIMP3 has been reported as a tumour suppressor in several human malignancies, with reduced expression correlating with poor clinical outcome. In this study, we investigated the changes in TIMP3 expression during melanoma progression. PATIENTS AND METHODS: TIMP3 expression was analysed by immunohistochemistry in sequential archived tumour material from stage I/II, stage III and stage IV samples from melanoma patients (n = 33). Protein expression was investigated for associations with disease-free survival and overall survival. Methylation status of the gene promoter was determined using methylation-specific PCR. In vitro assays were used to investigate the functional consequences of TIMP3 expression on behavioural aspects of melanoma cells. RESULTS: We show that TIMP3 expression decreases with melanoma progression although no significant clinical associations were obtained. Analysis of the status of promoter methylation using methylation-specific PCR revealed it to be a low-frequency event in melanoma. Additionally, through gene modulation experiments in melanoma cell lines, we show that TIMP3 negatively regulates cell migration, invasion and anoikis resistance. CONCLUSIONS: Collectively, our data suggests that TIMP3 functions as a tumour suppressor in melanoma and negatively regulates several aspects of the metastatic cascade.

Modeling and simulation of maintenance treatment in first-line non-small cell lung cancer with external validation.

BACKGROUND: Maintenance treatment (MTx) in responders following first-line treatment has been investigated and practiced for many cancers. Modeling and simulation may support interpretation of interim data and development decisions. We aimed to develop a modeling framework to simulate overall survival (OS) for MTx in NSCLC using tumor growth inhibition (TGI) data. METHODS: TGI metrics were estimated using longitudinal tumor size data from two Phase III first-line NSCLC studies evaluating bevacizumab and erlotinib as MTx in 1632 patients. Baseline prognostic factors and TGI metric estimates were assessed in multivariate parametric models to predict OS. The OS model was externally validated by simulating a third independent NSCLC study (n = 253) based on interim TGI data (up to progression-free survival database lock). The third study evaluated pemetrexed + bevacizumab vs. bevacizumab alone as MTx. RESULTS: Time-to-tumor-growth (TTG) was the best TGI metric to predict OS. TTG, baseline tumor size, ECOG score, Asian ethnicity, age, and gender were significant covariates in the final OS model. The OS model was qualified by simulating OS distributions and hazard ratios (HR) in the two studies used for model-building. Simulations of the third independent study based on interim TGI data showed that pemetrexed + bevacizumab MTx was unlikely to significantly prolong OS vs. bevacizumab alone given the current sample size (predicted HR: 0.81; 95 % prediction interval: 0.59-1.09). Predicted median OS was 17.3 months and 14.7 months in both arms, respectively. These simulations are consistent with the results of the final OS analysis published 2 years later (observed HR: 0.87; 95 % confidence interval: 0.63-1.21). Final observed median OS was 17.1 months and 13.2 months in both arms, respectively, consistent with our predictions. CONCLUSIONS: A robust TGI-OS model was developed for MTx in NSCLC. TTG captures treatment effect. The model successfully predicted the OS outcomes of an independent study based on interim TGI data and thus may facilitate trial simulation and interpretation of interim data. The model was built based on erlotinib data and externally validated using pemetrexed data, suggesting that TGI-OS models may be treatment-independent. The results supported the use of longitudinal tumor size and TTG as endpoints in early clinical oncology studies.

Phase 1 trial of vocimagene amiretrorepvec and 5-fluorocytosine for recurrent high-grade glioma.

Toca 511 (vocimagene amiretrorepvec) is an investigational nonlytic, retroviral replicating vector (RRV) that delivers a yeast cytosine deaminase, which converts subsequently administered courses of the investigational prodrug Toca FC (extended-release 5-fluorocytosine) into the antimetabolite 5-fluorouracil. Forty-five subjects with recurrent or progressive high-grade glioma were treated. The end points of this phase 1, open-label, ascending dose, multicenter trial included safety, efficacy, and molecular profiling; survival was compared to a matching subgroup from an external control. Overall survival for recurrent high-grade glioma was 13.6 months (95% confidence interval, 10.8 to 20.0) and was statistically improved relative to an external control (hazard ratio, 0.45; P = 0.003). Tumor samples from subjects surviving more than 52 weeks after Toca 511 delivery disproportionately displayed a survival-related mRNA expression signature, identifying a potential molecular signature that may correlate with treatment-related survival rather than being prognostic. Toca 511 and Toca FC show excellent tolerability, with RRV persisting in the tumor and RRV control systemically. The favorable assessment of Toca 511 and Toca FC supports confirmation in a randomized phase 2/3 trial (NCT02414165).

Increased PTP1B expression and phosphatase activity in colorectal cancer results in a more invasive phenotype and worse patient outcome.

Cell signaling is dependent on the balance between phosphorylation of proteins by kinases and dephosphorylation by phosphatases. This balance if often disrupted in colorectal cancer (CRC), leading to increased cell proliferation and invasion. For many years research has focused on the role of kinases as potential oncogenes in cancer, while phosphatases were commonly assumed to be tumor suppressive. However, this dogma is currently changing as phosphatases have also been shown to induce cancer growth. One of these phosphatases is protein tyrosine phosphatase 1B (PTP1B). Here we report that the expression of PTP1B is increased in colorectal cancer as compared to normal tissue, and that the intrinsic enzymatic activity of the protein is also enhanced. This suggests a role for PTP1B phosphatase activity in CRC formation and progression. Furthermore, we found that increased PTP1B expression is correlated to a worse patient survival and is an independent prognostic marker for overall survival and disease free survival. Knocking down PTP1B in CRC cell lines results in a less invasive phenotype with lower adhesion, migration and proliferation capabilities. Together, these results suggest that inhibition of PTP1B activity is a promising new target in the treatment of colorectal cancer and the prevention of metastasis.

Association of TIMP3 expression with vessel density, macrophage infiltration and prognosis in human malignant melanoma.

AIMS: Several anti-tumour properties have been ascribed to the tissue inhibitor of matrix metalloproteinases-3 (TIMP3) gene, including inhibition of neovascularisation in tumour xenografts. Reduced protein expression has been linked to promoter hypermethylation and allelic loss of heterozygosity in various human malignancies. In melanoma-positive lymph nodes from patients, we evaluated the association between TIMP3 expression, vessel density, macrophage infiltration and potential correlations with disease-free survival (DFS) and overall survival (OS). PATIENTS AND METHODS: TIMP3 expression was analysed by immunohistochemistry (IHC) in melanoma lymph node biopsies of stage III melanoma patients (n = 43). Blood vessel density and macrophage infiltration were quantitatively assessed and correlation with TIMP3 expression was investigated. Methylation status of the gene promoter was determined using methylation-specific polymerase chain reaction (MSP). Protein expression and promoter methylation status were investigated for associations with DFS and OS. RESULTS: Reduced expression of TIMP3, as determined by IHC, was observed in 74% of the cases (32 in 43). A significant inverse correlation was observed between TIMP3 expression and vessel density (p = 0.031). Correlation between TIMP3 expression and macrophage infiltration was not statistically significant (p = 0.369). MSP analysis revealed methylation of the gene promoter in 18% (7 in 38) of the analysed cases. No differences in OS and DFS were observed between cases with high and low TIMP3 expression. Gene promoter methylation was significantly associated with both poor 5-year DFS (p = 0.024) and OS (p = 0.034). CONCLUSIONS: Our data indicate that TIMP3 is a dominant negative regulator of angiogenesis in cutaneous melanoma and gene silencing by promoter methylation is associated with poor outcome.

A ring barrier-based migration assay to assess cell migration in vitro.

Cell migration is a key feature of virtually every biological process, and it can be studied in a variety of ways. Here we outline a protocol for the in vitro study of cell migration using a ring barrier-based assay. A 'barrier' is inserted in the culture chamber, which prevents cells from entering a defined area. Cells of interest are seeded around this barrier, and after the formation of a peripheral monolayer the barrier is removed and migration into the cell-free area is monitored. This assay is highly reproducible and convenient to perform, and it allows the deduction of several parameters of migration, including total and effective migration, velocity and cell polarization. An advantage of this assay over the conventional scratch assay is that the cells move over an unaltered and virgin surface, and thus the effect of matrix components on cell migration can be studied. In addition, the cells are not harmed at the onset of the assay. Through computer automation, four individual barrier assays can be monitored at the same time. The procedure can be used in a 12-well standard plate allowing higher throughput, or it can be modified to perform invasion assays. The basic procedure takes 2-3 d to complete.

Low molecular weight protein tyrosine phosphatase (LMWPTP) upregulation mediates malignant potential in colorectal cancer.

Phosphatases have long been regarded as tumor suppressors, however there is emerging evidence for a tumor initiating role for some phosphatases in several forms of cancer. Low Molecular Weight Protein Tyrosine Phosphatase (LMWPTP; acid phosphatase 1 [ACP1]) is an 18 kDa enzyme that influences the phosphorylation of signaling pathway mediators involved in cancer and is thus postulated to be a tumor-promoting enzyme, but neither unequivocal clinical evidence nor convincing mechanistic actions for a role of LMWPTP have been identified. In the present study, we show that LMWPTP expression is not only significantly increased in colorectal cancer (CRC), but also follows a step-wise increase in different levels of dysplasia. Chemical inhibition of LMWPTP significantly reduces CRC growth. Furthermore, downregulation of LMWPTP in CRC leads to a reduced migration ability in both 2D- and 3D-migration assays, and sensitizes tumor cells to the chemotherapeutic agent 5-FU. In conclusion, this study shows that LMWPTP is not only overexpressed in colorectal cancer, but it is correlated with the malignant potential of this cancer, suggesting that this phosphatase may act as a predictive biomaker of CRC stage and represents a rational novel target in the treatment of this disease.

Incidence and relevance of proteinuria in bevacizumab-treated patients: pooled analysis from randomized controlled trials.

BACKGROUND: This analysis evaluated the incidence of and risk factors for bevacizumab-related proteinuria and assessed for any associated clinical sequelae, including renal function changes. METHODS: Patient-level adverse event and laboratory data from a pooled safety database were used to characterize alterations in urine protein excretion following interventional therapy ± bevacizumab in 17 randomized trials across multiple tumor types. Severity of renal function change was assessed using changes in serum creatinine concentration from baseline values. Potential predictors of proteinuria and the association between proteinuria and other adverse events were also investigated. RESULTS: Among 14,548 patients, the incidence of any-grade proteinuria was 8.2% (733/8,917) and 4.6% (257/5,631) in the bevacizumab and control groups, respectively; rates of grade ≥3 proteinuria were 1.4 and 0.2%. Post-baseline proteinuria grade and bevacizumab were associated with increased rates of renal dysfunction. Patients developing proteinuria had an increased rate of any-grade infection but not thromboembolic events. History of diabetes was the only examined risk factor that appeared to have a significant association with proteinuria development. CONCLUSIONS: This analysis confirmed a significant increase in the development of proteinuria during bevacizumab treatment. We also observed an increased rate of renal dysfunction associated with bevacizumab treatment and among subjects with proteinuria, although the dysfunction was generally mild. The development of proteinuria was also associated with a modest increase in risk of infection.

The use of phase 2 interim analysis to expedite drug development decisions.

PURPOSE: To expedite drug development, we propose a two-step decision-making process that utilizes interim efficacy results from a comparative phase 2 trial to determine whether to accelerate subsequent phase 3 preparations, and final analysis to ultimately determine whether to conduct phase 3 testing. METHODS: The operational characteristics of this process were evaluated by modeling simulated data of oncology trials and retrospectively analyzing data from historical comparative phase 2 trials. Progression-free survival (PFS) was used as the primary endpoint; the estimated PFS hazard ratios (HRs) of ≤0.60 at interim and of ≤0.65 at final analysis favoring the experimental arm were defined as positive results. The conditional probability of achieving a target PFS HR at final analysis, based on observed interim results, was also estimated by imputing post-interim data with and without the proportional hazard assumption. RESULTS: Simulations of phase 2 trials showed that estimated interim PFS HRs correlated with estimated final PFS HRs, with reasonably low false-positive rates for supporting phase 3 "go" decisions at interim. Using observed historical data, decisions based on interim PFS analyses also matched final phase 3 "go" and "no-go" decisions with a false-positive rate of 16.7% (2 of 12 trials) and a false-negative rate of 9.4% (3 of 32 trials). Analytical modeling accurately predicted final PFS HRs from observed interim data when accounting for appropriate underlying assumptions. CONCLUSIONS: The results support the usefulness of a two-step decision-making process that utilizes interim phase 2 results to reduce the interval between phase 2 completion and phase 3 initiation.

Differential TIMP3 expression affects tumor progression and angiogenesis in melanomas through regulation of directionally persistent endothelial cell migration.

The angiogenic potential of solid tumors, or the ability to initiate neovasculature development from pre-existing host vessels, is facilitated by soluble factors secreted by tumor cells and involves breaching of extracellular matrix barriers, endothelial cell (EC) proliferation, migration and reassembly. We evaluated the angiogenic potential of human melanoma cell lines differing in their degree of aggressiveness, based on their ability to regulate directionally persistent EC migration. We observed that conditioned-medium (CM) of the aggressive melanoma cell line BLM induced a high effective migratory response in ECs, while CMs of Mel57 and 1F6 had an inhibitory effect. Further, the melanoma cell lines exhibited a varied expression profile of tissue inhibitor of metalloproteinase-3 (TIMP3), detectable in the CM. TIMP3 expression inversely correlated with aggressiveness of the melanoma cell line, and ability of the respective CMs to induce directed EC migration. Interestingly, TIMP3 expression was found to be silenced in the BLM cell line, concurrent with its role as a tumor suppressor. Treatment with recombinant human TIMP3 and CM of modified, TIMP3 expressing, BLM cells mitigated directional EC migration, while CM of TIMP3 silenced 1F6 cells induced directed EC migration. The functional implication of TIMP3 expression on tumor growth and angiogenic potential in melanoma was evaluated in vivo. We observed that TIMP3 expression reduced tumor growth, angiogenesis and macrophage infiltration of BLM tumors while silencing TIMP3 increased tumor growth and angiogenesis of 1F6 tumors. Taken together, our results demonstrate that TIMP3 expression correlates with inhibition of directionally persistent EC migration and adversely affects the angiogenic potential and growth of melanomas.

Progression-free survival as a surrogate endpoint for overall survival in glioblastoma: a literature-based meta-analysis from 91 trials.

BACKGROUND: The aim of this study was to determine correlations between progression-free survival (PFS) and the objective response rate (ORR) with overall survival (OS) in glioblastoma and to evaluate their potential use as surrogates for OS. METHOD: Published glioblastoma trials reporting OS and ORR and/or PFS with sufficient detail were included in correlative analyses using weighted linear regression. RESULTS: Of 274 published unique glioblastoma trials, 91 were included. PFS and OS hazard ratios were strongly correlated; R(2) = 0.92 (95% confidence interval [CI], 0.71-0.99). Linear regression determined that a 10% PFS risk reduction would yield an 8.1% ± 0.8% OS risk reduction. R(2) between median PFS and median OS was 0.70 (95% CI, 0.59-0.79), with a higher value in trials using Response Assessment in Neuro-Oncology (RANO; R(2) = 0.96, n = 8) versus Macdonald criteria (R(2) = 0.70; n = 83). No significant differences were demonstrated between temozolomide- and bevacizumab-containing regimens (P = .10) or between trials using RANO and Macdonald criteria (P = .49). The regression line slope between median PFS and OS was significantly higher in newly diagnosed versus recurrent disease (0.58 vs 0.35, P = .04). R(2) for 6-month PFS with 1-year OS and median OS were 0.60 (95% CI, 0.37-0.77) and 0.64 (95% CI, 0.42-0.77), respectively. Objective response rate and OS were poorly correlated (R(2) = 0.22). CONCLUSION: In glioblastoma, PFS and OS are strongly correlated, indicating that PFS may be an appropriate surrogate for OS. Compared with OS, PFS offers earlier assessment and higher statistical power at the time of analysis.