Components of a competitive Letter of Intent (LOI)
As with any application for support, a great idea can be poorly communicated or well communicated. The review criteria for LOI are:
1) Strong scientific hypothesis
2) Supporting preliminary data and/or a strong rationale
3) Adequate patient accrual
4) Innovative and well-justified correlative studies
5) Ability to meet regulatory requirements
6) Not duplicative
7) Agent availability
Industry sponsor concurrence
LOIs with great ideas (or even good ideas), which do not provide specific information about the rationale/ hypothesis, supportive data, trial design, patient population, treatment plan and statistical evaluation generally are assigned worse scores by CTEP than they would receive if that information had been provided. Competitive LOIs most effectively assure CTEP that the question being asked is important and that the methodology applied to the question will be sound. While all components of LOIs are reviewed, the sections discussed below are critical to the CTEP review process.
The rationale for performing the specific clinical trial should be explicitly stated. There is no need to reproduce data from either the solicitation for the LOI or general medical literature on the disease. However, a specific reason to test an agent should be offered. Reasons might include compelling in vitro or in vivo experiments, molecular biological rationales (drug “hits“ target protein, and this protein is selectively over-expressed in the tumor type to be studied in the LOI.), or prior clinical data such as a significant number of tumor responses in prior studies using the agent or similar agents.
The hypothesis should be succinctly stated as a question to be asked as the primary endpoint of the study, and the design of the study should permit an unequivocal answer to the question.
A poorly stated rationale and hypothesis: “Cancer X has a poor prognosis and there is no approved effective therapy. It is possible that agent Y will demonstrate activity against cancer X in humans.”
A well-stated rationale and hypothesis: “Cancer X is known to over express the Q receptor in 75% of cancer X specimens sampled from patients who have a recurrence after definitive regional therapy. Drug Y binds to and inactivates the growth stimulating effects of receptor Q, and in multiple animal tumor models of cancer X, drug Y has been shown to have twice the tumor shrinkage rate as commercially available drugs, and cures 30% of all mice treated. We hypothesize that drug Y will demonstrate an overall response rate of at least 30% by RECIST criteria when administered to chemotherapy na´ve patients with cancer X, whose cancers have recurred following regional therapy, compared to 15% who historically respond when treated with commercially available drugs.
Supportive data: If you have used this agent in your laboratory or in a clinical trial at your site, and have unpublished data that is relevant, include a summary of that data, with figures, pre-prints or a summary from a grant progress report or grant application, if available, for the reviewers. It is helpful to receive data such as: “of 8 patients with prostate cancer in the phase I trial, who had disease progression after prior therapy with docetaxel, 4 receiving docetaxel with agent Y had PR sustained from 3-9 months”, compared to: “we saw some activity in our (unpublished) phase I trial that we want to pursue in this phase II trial”
2. Endpoints/Statistical Considerations: The primary endpoint must be explicitly stated, and the statistics supporting the endpoint and trial design must be provided.
A poorly-stated endpoint: We will assess progression free survival (PFS), toxicity, serum protein X levels and QOL. We will declare a PFS of 35% to be interesting.
A well- stated endpoint: The primary end point will be 6 month progression free survival (PFS). Secondary exploratory evaluations of toxicity and QOL, and serum protein X will be made. Based on our institutions prior 5 trials including 200 patients, the proportion of patients alive without progression following standard therapy is 15% (95% c.i. 10-19%). Therefore, the null hypothesis is that 15% will remain alive without progression at 6 months, and our hypothesis is that use of drug Y will increase the PFS to 35%. A two-stage design will be used. If at least 4/20 patients achieve a PFS of 6 months, a total of 44 patients will be accrued, assuming a 10% ineligibility rate. If 11/40 evaluable patients achieve a PFS of 6 months, we will conclude agent Y should be further explored in this setting.
3. Laboratory correlates: A discussion concerning the rationale, hypothesis, and methods for laboratory correlates should be succinct, yet specific enough for CTEP to evaluate the relative merits of the correlates. The choice of targets to be evaluated and the choice of assays that will be used to assess them should be explicitly justified, and reasons for not looking at other targets/using other assays should be included.
A poorly stated correlatives section: We will look at Q, R, S, T, U, and V receptor levels and circulating t*, u* and v* from whole blood, and correlate these levels with outcome.
A well-stated correlatives section: As noted above, drug Y binds to and inactivates the growth stimulating effects of receptor Q. This presence of the activated target (expected in 75% of patients with X cancer) can be evaluated with immunohistochemistry methods (ref). We will measure Q receptor levels on paraffin embedded tumor samples from biopsies taken within 2 months prior to enrollment on this trial. Commercial “Q-assay” kits will be used or, pending funding, the assay method of Block and Cube [citation] utilizing standard reagents will be used. There is no standard cutoff between normal and abnormal Q receptor levels. We will use the median Q receptor levels defined in this population to separate low versus high expressors, and using the statistical technique of Sphere and Globe [citation] we will look for an association between baseline tumor Q receptor expression and response rate. Although there has been some suggestion that T and V, downstream of Q, might be related to Q inhibition by Y, it has not been possible to measure phosphoT or phosphoV reliably in specimens from murine tumors, and there is no reason to think that modulation of Q will alter the total T and total V in cells.
4. Patient population: In the LOI, you have the opportunity to note prior studies in this patient population and your accrual rate. If you project some very different and improved accrual rate, state why: e.g. Although our last trial in AML accrued only 5 patients in 3 years from 1998-2001, we have since recruited Drs. X, Y, and Z, who together have seen 200 potentially eligible new patients in the last 12 months, and were responsible for consenting 45, 62, and 93 patients per year for clinical trials before they left St. Elsewhere. Thus, we anticipate being able to accrue 4 patients/month to this trial, despite our pathetic prior track record.
5. Competing studies: Please note when you expect competing trials to be completed, so that we can discount the problem of competing studies. Having 3 open competing studies in the same patient population doesn't give reviewers great confidence that this trial will receive priority and have adequate accrual.
We cannot stress the value of clarity and brevity enough. Most accepted LOIs are in the 4-6 page range. Exceedingly long LOIs distract the reviewers from the essence of the proposal. Fully written protocols submitted as LOIs are unlikely to earn a score higher than what would have been assigned to a much shorter well-written LOI.[Author ID1: at Tue Apr 29 16:43:00 2003 ]
Created on 5/2/03 12:18 PM, I:\Components of a competitive Letter of Intent draft.doc A D Colevas
Page 1 of 3