Who wants to go first? A simulation study of accrual in a stand-alone trial versus starting a platform trial.

Others have quantified the efficiency of the platform approach as compared to a sequence of independent two-arm trials and have shown the platform approach more efficiently evaluates a set of candidate therapies. However, a practical barrier to initiating a platform trial is incentivizing the first candidate therapies to enter the platform. A platform trial is more complex and will take longer to design and operationalize than a traditional trial. For the first therapy, this additional up-front planning time must be considered along with the ability to enroll. There is a common concern that accrual in a platform setting would take longer than for a single stand-alone trial because intuition suggests that a two-arm trial with a smaller total sample size should complete accrual more quickly than a multi-armed trial. We focus on the accrual duration for the first therapy as a particular barrier to initiating a platform trial strategy. We simulate accrual into a platform trial versus a stand-alone trial in the setting of a large clinical trial network. Accrual duration in the platform strategy dominates that of a single stand-alone trial if the platform leverages a large enough fraction of the site network. Patient preference for a particular stand-alone trial has little impact on the ability of a stand-alone trial to enroll more quickly.

Publisher Correction: ASPI: a public-private partnership to develop treatments for autism.

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

Cyclooxygenase-2 expression in human breast cancers and adjacent ductal carcinoma in situ.

Cyclooxygenase-2 (COX-2) is an inducible enzyme that converts arachidonic acid to prostaglandins. Overexpression of the COX-2 gene in mammary glands of transgenic mice was sufficient to induce tumorigenesis. We analyzed COX-2 expression in human breast cancers (and breast cancer cell lines) and adjacent ductal carcinoma in situ (DCIS) as well as its association with HER2/neu and clinicopathological variables. Archival primary breast carcinomas (n = 57), adjacent DCIS (n = 14) and DCIS alone (n = 2) were analyzed for COX-2 and HER2 expression by immunohistochemistry using specific monoclonal antibodies. An immunohistochemical scoring system was used. HER2 gene amplification had been analyzed previously by fluorescence in situ hybridization (n = 20). Histology of carcinomas included infiltrating ductal (n = 44), lobular (n = 2), and other (n = 7). Frozen breast cancers and adjacent normal tissue pairs (n = 9) were analyzed for COX-2 mRNA by reverse transcription-PCR. COX-2 and HER2 expression were also analyzed in human breast cancer cell lines (MCF-7, MCF-7/HER2, SK-BR-3, and MDA-MB-231) by immunoblotting. Cytoplasmic COX-2 expression was detected at an intermediate or high level in epithelial cells in 18 of 42 (43%) invasive breast cancers and in 10 of 16 (63%) cases of DCIS. Normal-appearing breast epithelia adjacent to cancer expressed COX-2 in 81% of cases and was generally focal and of similar or decreased intensity relative to adjacent neoplastic epithelia. COX-2 mRNA was detected in all samples analyzed by reverse transcription-PCR and was increased in eight of nine breast cancers relative to paired normal tissue. In archival tumors, no significant correlation was found between COX-2 and HER2 expression/amplification and clinicopathological variables. COX-2 expression was induced in MCF-7 cells stably transfected with HER2, in contrast to parental MCF-7 cells, and was detected in MDA-MB-231, but not SK-BR-3 cells. COX-2 is frequently overexpressed in invasive breast cancers and in adjacent DCIS and, thus, may be an early event in mammary tumorigenesis. Forced HER2 expression in MCF-7 cells was shown to up-regulate COX-2, although no association was found in human tumors. Our results suggest that nonsteroidal anti-inflammatory drugs and selective COX-2 inhibitors may be useful in the chemoprevention and therapy of human breast cancer.

Suppression of tumorigenesis and induction of p15(ink4b) by Smad4/DPC4 in human pancreatic cancer cells.

PURPOSE: The tumor suppressor gene Smad4/DPC4, a key transcription factorin transforming growth factor beta (TGF-beta) signaling cascades,is inactivated in 50% of pancreatic adenocarcinomas. We seek to determine the role of Smad4/DPC4 in the suppression of tumor cell growth and in the regulation of TGF-beta-mediated expression of cell-cycle regulatory genes p15(ink4b) and p21(waf1). EXPERIMENTAL DESIGN: Smad4/DPC4 is overexpressed by adenoviral infection in CFPac-1 pancreatic cancer cells, in which the Smad4/DPC4 is homozygously deleted, and in Capan-1 pancreatic cancer cells, in which Smad4/DPC4 is not expressed. Expression of the TGF-beta downstream target gene p21(waf1), regulation of the p15(ink4b) promoter, anchorage-independent growth, and tumorigenesis were examined. RESULTS: We demonstrate that expression of Smad4/DPC4 in Capan-1 cells reduced anchorage-independent growth by more than 50%, and inhibited xenograft tumor growth. However, overexpression of Smad4/DPC4 did not inhibit CFPac-1 cell growth. Interestingly, Smad4/DPC4 induced expression of p15(ink4b), p21(waf1), and TGF-beta-responsive reporter gene in Capan-1 but not in CFPac-1 cells. Furthermore, we found a previously unidentified Smad4 binding element (SBE) located in the region between -356 and -329 bp of the p15(ink4b) promoter. The p15(ink4b) promoter reporter gene assays revealed that Smad4-dependent transcriptional activation is mediated by this SBE, which indicates that p15(ink4b) is one of the downstream target genes regulated by Smad/DPC4. CONCLUSION: These results explain the role of Smad4/DPC4 in TGF-beta-mediated inhibition of cell proliferation in vitro and in vivo. Moreover, these results suggest that Smad4/DPC4-mediated tumor suppression and induction of TGF-beta-regulated cell-cycle-inhibitory genes may depend on additional factors that are absent in CFPac-1 cells.

PAX6 suppresses growth of human glioblastoma cells.

PURPOSE: Glioblastomas (GBMs) are the most common primary malignant brain tumors. Majority of GBMs has loss of heterozygosity of chromosome 10. The PAX6 encodes a transcription factor that involves in development of the brain, where its expression persists. We have reported that the expression of PAX6 was significantly reduced in GBMs and that a low level of PAX6 expression is a harbinger of an unfavorable prognosis for patients with malignant astrocytic glioma. Interestingly, PAX6 expression was increased in suppressed somatic cell hybrids derived from introducing a normal human chromosome 10 into U251 GBM cells. Thus it is interesting to determine if repression of PAX6 expression is involved in anti-tumor suppression function in GBM. EXPERIMENTAL DESIGN: We overexpressed PAX6 in a GBM cell line U251HF via either stable transfection or infection with recombinant adenovirus, and examined cell growth in vitro and in vivo. RESULT: Although we did not observe changes in the cell doubling time for PAX6-stable transfectants, significantly fewer numbers of PAX6-positive colonies grew in soft agar. Transient overexpression of PAX6 via adenovirus, however, suppressed cell growth by increasing the number of cells in G1 and by decreasing the number of cells in S-phase, and later on caused a dramatic level of cell death. Repeated subcutaneous and intracranial implantation experiments in nude mice using PAX6-stable transfectants provided solid evidence that PAX6 suppressed tumor growth in vivo and significantly extended mouse survival. CONCLUSION: Our data demonstrate that PAX6exerts a tumor suppressor function that limits the growth of GBM cells.