 Thanks for having me here today. I'm a pathologist. I've been at the FDA in OIR for three years And today I'm going to try to help you understand When you need an IDE? How do you know when you need an IDE? I think the rules for that are relatively simple Sometimes they're easy to apply. Sometimes they're not and we'll get into that so from our Code of Federal Regulations a significant risk device is defined as an Investigational device that and we for in vitro diagnostic devices. We use really parts three and four of this definition So it's for use of substantial importance in diagnosing curing mitigating or treating disease or otherwise preventing impairment of Human health and presents a potential for serious risk to the health safety or welfare of a subject Or otherwise presents a potential for serious risk to the health safety or welfare of the subject now There are different types of devices and We want to make a clear distinction here between in vitro diagnostic devices Which is what we deal with and what we're talking about today versus implanted devices Now in an implanted device, it's very simple if you have a prosthetic heart valve The risk of the trial is the same as the risk of the device. It's very easy It's different for in vitro diagnostic devices the risk of the trial depends on how the information generated by the device is used so there's a Distinction that doesn't really have to be made between the device and the trial because the determination is really based on both The risk determination question for in vitro diagnostic devices is what is the risk of the use of this device in this specific trial? So the use of the identical device one device can be either significant risk or non-significant risk Depending on how it's used in the trial and you can have two trials using the same device One would be an SR trial. The other would be an NSR trial I want to bring up something that dr. Khan raised earlier. He mentioned a device master file If you show us a device and you give us all possible information about it We can't tell you if it's SR or NSR unless you tell us how it's being used in a trial But if you're using a single device in multiple trials You can make it easier on yourself and on us if you submit a device master file and then trial by trial if you need an IDE for a trial you can Send in that IDE and reference the the analytical validation data in the master file That's just one way of sort of simplifying the process if you're in that situation Now, I'm a pathologist as I said I deal mainly with companion diagnostic devices for oncology trials I'm not a molecular biologist But the general principles that I'm going to give you any oncology and drug examples The same general principles apply to genomics devices We use the same rules There may be some Finer points that some of the other people today can get into But I'm going to give you the basic rules that we use in the oncology trial and there are basically four questions Four key questions with respect to risk First will the results from the device be used for enrollment? Second will the results from the device cause patients to forego known effective or approved therapy Third will the results from the device cause patients to be subjected to unacceptable Toxicities and fourth will the results from the device cause patients to undergo a pet potentially high-risk biopsy There's one additional question that we usually don't have to address because there's usually not enough data from drug trials, but sometimes there is and Based on available data. Is there a known biomarker effect with respect to either toxicity or effectiveness? If there is a known effect then the use of the device could Be of risk by increasing toxicity or diverting a patient for more effective therapy Well, there are certain things that we don't consider at all in a risk determination. We don't consider benefit We only look at risk now as David indicated earlier in the evaluation of an IDE We do look at risk and benefit But the question of whether you need an IDE a potential benefit is not part of that equation. It's only risk Another thing we don't consider is the number of patients at risk If you have a trial with three patients in it It's not NSR just because you have three patients that we We undergo the same decision-making process regardless of the number of patients in the trial Incorrect exclusion from a trial Would not be considered a risk because we would expect management of that patient to revert to standard of care So what do you need for risk determination for an in vitro diagnostic device? Well, we like to see a complete clinical trial protocol This would of course include the inclusion and exclusion criteria Proposed interventions based on the device. How is the in vitro diagnostic device used to make decisions about treatment? or enrollment We'd like a general understanding of the device and sampling requirements Now I have a black box here and I use that black box intentionally When I downloaded this presentation last night onto my iPhone this box was green. I'm glad that it's it's black right now But you know, it's a black box because we you know, the truth is we don't really look much at the details of the device In this risk determination. We look at the device output And how the information is being used so very simply a device will give you a Say a marker negative or a marker positive result If that result is used for enrollment or arm assignment or stratification or other purposes that will determine the level of risk and whether the Study is significant or non-significant risk The most the simplest Example of an NSR trial is a device where the output is used purely for exploratory purposes the information is not being used to drive anything and Is not being used to put patients in a particular arm of the trial or drive treatment So if it's purely for explore exploratory Most likely a trial like that would be an NSR trial Now if the device output is used as a criterion of enrollment Then the device exposes the patient to all the risks of being in the trial and these risks could include Forgoing known effective therapy drug toxicities or new biopsies performed solely for trial purposes a Trial might be stratified using known or suspected prognostic indicators For example age gender or smoking history and if the device output is used purely for stratification within trial arms Then this would not change the risks to which the patient is exposed and that would generally Be an NSR trial however Sometimes You have for example a trial where subjects are accrued without respect to device results Until let's say 50 marker negatives are enrolled and subsequent to that point only marker positives will be enrolled Well that the determination with that would be the same as a marker based enrollment that and that would determine that the risk of that would Be determined by what happens to the patient based on the test So let's say you have a trial where patient everybody's enrolled But arm a Put marker positives and arm be you have marker negatives. Well, this would be an NSR trial But only if there's equipoise ie if the arms are both equal or of unknown risk And that's sometimes true, but it's often not true for example, if you have arm a and arm be where arm a is an experimental agent and R&B's Patients get standard of care treatment which is known to be effective Anybody who's enrolled in arm a is going to forego known effective therapy Therefore an incorrect test result will potentially deprive a patient of alternative effective therapy and this would generally be classified as a significant risk trial Yeah, the experimental agent might prove to be better than the standard of care, but it might be worse We just don't know so it's it's clearly a risk Now there's a situation where sometimes everybody in a trial gets standard of care But the experimental agent is an add-on to standard of care And in many situations we we consider this non-significant risk because everybody is getting standard of care But there is an exception to this With regard to drug trials There are some situations where there's usually not a lot of information available about toxicity of experimental drugs, but sometimes we have enough information to know that That an agent is particularly toxic or more toxic than what the patient would have otherwise received so An arm a study where everybody gets standard of care, but in one arm the experimental agent has expected severe toxicity if the Arms are divided by a marker positive marker negative in other words if the test Determines which arm the patients go into this would be an SR trial Because the patients Would be going Undergoing added toxicity beyond that expected with standard of care agents Okay, now I've used that phrase standard of care. We could probably Talk all day long about what is standard of care? we In oncology and drug trials and companion diagnostic devices for Those trials we find NCCN guidelines very useful They're not carved in stone. We know that standard of care does change, but they are used useful and we use them frequently There are two things I'd like to point out about NCCN NCCN guidelines They do use the phrase clinical trial In in various parts of the guidelines Now you'll note that on the bottom of every page of every guideline in every edition You will see this statement clinical trials and CC believe NCCN believes that the best management of any cancer patient And it is in a clinical trial Participation in clinical trials is especially encouraged Well, that's fine, and I'm not sure if everybody agrees with this, but this does not affect our risk determination This Does not mean the trial all trials are safe or there is no risk However elsewhere in the NCCN guidelines appear for example for patients with renal cell carcinoma if they have stage 4 surgically unrespectable renal cell carcinoma First-line therapy there are several options one of which is clinical trial and we take this to mean that That there really is no really good therapy that the patient is going to forego so So So that again, that's that's sort of one of the ways that we use the NCCN guidelines Okay biopsy risk this this is sometimes considered a and can be a complicated question But it's important to recognize that If enrolled patients in a trial will undergo a biopsy beyond what would be considered standard of care For the sole purpose of development of the test of the device Then the risks of the biopsy are attributed to the device So accordingly a high-risk biopsy in this setting would generally warrant a significant risk determination now Every patient is different and biopsy risk varies widely and it depends on a lot of factors These include the site of the procedure the type of the procedure patients disease and underlying health and Institutional experience and support capabilities in the in the context of a clinical trial The biopsy risk is controlled according to the clinical judgment of the health care providers Now just as examples of potentially high-risk biopsies. There's lung, medial stynum, brain, pancreas These are generally sites that you don't want a biopsy unless you really have to Examples of biopsies that are likely low-risk are for example skin endoscopic gastrointestinal biopsies and cervix Now we acknowledge that in studies It means oncology studies of recurrent and end stage cancers the site that will need to be biopsied May not be known in advance and it generally is not known in advance when the protocol is written But it's important to keep in mind that a patient on a clinical trial Using an investigational device should not undergo about high-risk biopsy solely device for device development unless there's an approved IDE so In summary It's not the device that's high-risk. It's not the trial that's high-risk It's it depends on the specific use of the device in a specific trial and Again the output of the device How the output of the device is used on up in the trial. Thank you