Do your own DD. I hate it when people purchase a stock without knowing a damn thing about it.
Understand what this company does and where it currently is. Research what AppyScore is and why it went from $7.50 to $1.50. FDA told them to revise their study and instead of 600 patients to increase to 800. The current stage APPY is at is data analysis. After the data has been presented (within a week or so) they will file again with the FDA. They still expect the detection rate to be about ~94%.
There are 37.5 million shares out, meaning a small float. They have enough cash (about 20 mil) and their current cash burn rate is about 1.3mil every quarter.
Easily, I can say this is a $6 stock without FDA filing news. After the data and FDA filing, $10. Once approved, it could be in $30s
Another benefit of Appy Score with the cassette reader is that the results will be back in 15 minutes. No signs of an Appendicitis, ok you can go home now. Versus 2-3 hrs. of waiting for the CT reulsts. This makes the emergency room run much more effeciently. Time is money. JMHO
No way we will get 95% in this new trial in the primary end point from 89%. Nothing significant have change it enough to achieve 95%. We will see a sell off after results are release.
Nobody's saying they will get 95% magically (unless something is tweaked by the removal of the small irritants that were in the 600 patient trial).
I think the only person who has banked on the APPY standalone (though even he is aware of the combination tests) is Adam Feuerstein. He throws out a "got to be better than 89%", then he goes further and wants greater than 95%.
Obviously you cannot do that with APPYScore standalone - and WHY do it that way, if you have the combination tests which EASILY beat these numbers ?
Biotech Stock Mailbag: AspenBio
04/30/10 - 06:00 AM EDT
When AspenBio loosened the test's sensitivity in a post-hoc analysis of the initial phase III study, the negative predictive value rose to 89%. When AspenBio looked at data combining AppyScore and a standard test measuring white blood cell counts, the NPV rose to 93% (or 98% with loosened sensitivity.)
I'm throwing out a lot of NPV values here, but they're important to keep in mind as a gauge for what constitutes positive data from the new phase III study.
McGonegal didn't want to set expectations for the AppyScore NPV in the new phase III study, but I think it's very safe to assume that results from the new study have to improve on the 89% NPV recorded the last time around. Given the changes made to the new study, there is no reason why AppyScore shouldn't perform better.
What NPV score gets AppyScore approved by FDA? Again, McGonegal and Pusey don't have a specific number to throw out, saying that FDA has not set a specific efficacy hurdle for the test.
If I had to make an educated guess, AppyScore's new NPV must exceed 95%, and may even need to reach 98-99%, in order for 1) FDA to approve; and/or 2) for AppyScore to have any chance of being clinically relevant and therefore used by doctors in hospital emergency departments.
Dan Weiss seems better aware of the value of the combination tests - as he doesn't even mention what he wants the standalone test to be - but focuses on the combination test:
Preview of AspenBio's Potentially Blockbuster Clinical Trials 5 comments
April 25, 2010 | about: APPY / DNDN / HGSI / NVS
The critical level to look for especially with the AppyScore in combination with
either the white blood count or neutrophil count is a NPV of 90% or greater
which all indications from prior clinical results show should be achievable (see
my prior posts on AspenBio for more details). Strong clinical results from the
trial would provide the company with a very high likelihood of FDA approval
for the faster rapid test (expected in late 2010 or early 2011) as the key issue
for the FDA is efficacy as safety should not be a major concern for a blood
I understand they have better tracking this time around, but what is there to guarantee better tracking of patient can help results instead of making it worst? Also, they lost like 16-20 patients out of 600 from the previous test? how is accounting for 16-20 more patients can alter the results from 89% and jump to 95%? It's only like 3% of the patient trial size...
Yes, as I indicated above, I am not sure if better followup helps or hurts.
However your data is incorrect.
In the "600 patient trial" they had 586 patients and 220 excluded (most of which due to followup).
I don't know where you are getting your 16-20 patients from.
In the new 800 patient trial, APPY said they were getting much fewer losses to followup (in low teens).
The huge number of losses to followup in the 600 patient trial may ALSO have been ONE of the reasons APPY decided to do a new trial (see the "RED FLAG" comment below),
From the Feb 9, 2010 interim analysis conference audio (taken from notes - so not exact):
just one more quick question before I jump back in queue
can you give us a sense maybe on .. number of lost patients on followup
just qualitatively .. in previous study close to 800 patients and only 600 analyzed
was this an improvement over that .. or similar ?
unbelievable improvement ..
I don't want to give the number .. I happen to know it, because that is something we are not blinded to .. that it doesn't give us any data ..
I don't know exactly what the number will be at the end .. but it's a .. very small number of patients
and that's a good example of where we have done to improve the .. out procedures, our followup, our monitoring
the way we go to find patients if they don't come back
it's in very low digits .. which is .. OUTSTANDING for this type of study
ESPECIALLY a study that .. emergency department .. because you can imagine, many patients come in there
and don't even give full information - it's hard to track them down, so ..
so I think we've turned something that probably was a RED FLAG in the previous trial .. is something that we're doing exceptionally well ..
There is no way the first trial be 83% NPV then modified cutoff to get 89% NPV.
I assume you are talking about the prelim data PR which had data for both APPYScore standalone and combination (for cut-off of 15 and 20).
You don't believe that NPV can fall from 89% to 83% as you increase cut-off from 15 to 20 ?
The reason for that is that as you increase cut-off from 15 to 20 (and you look at patients who are scoring on APPYScore below cut-off - to be sent home as "non-appendicitis") you have more normal patients sent home, but you have even MORE appendicitis patients sent home (because distribution of appendicitis patients is denser at higher APPYScore scores). With the result that the ratio of normal patients sent home to (normal + appendicitis) patients sent home DECREASES.
For example if you have 10 normal patients below 15 and 1 appendicitis patient below 15.
NPV = 10/(10 + 1) = 0.909 i.e. 90.9%
If you increase cut-off from 15 to 20 and normal patients increases to 20, but appendicitis patients rapidly increases to 3, then:
NPV = 20/(20 + 3) = 0.860 i.e. 86.9%
Basically normal patients went up by 2x, but appendicitis by 3x (even though are still small i.e. only 3).
you gotta be kidding me!! standalone APPYScore data is the PRIMARY ENDPOINT. It is the most important result FDA considers!! and you said its irrelevant?! you are a funny man. well at least we both agree that standalone APPYScore results is gonna be short of expectation....
It is not short of expectations, but that one is not expecting anything of it - except be similar or better than before.
We already have APPYScore + WBC combination test which gives good results - as demoed by 600 patient trial.
It was problematic using THAT since there was a "post-hoc" smell to it earlier.
Therefore you have it designed-in into this new 800 patient trial - which is in fact a LEGITIMIZING of the combination test results.
Maybe the confusion is from this whole "primary endpoint" and "secondary endpoint" thing:
APPYScore standalone - primary endpoint
APPYScore + WBC - secondary endpoint
APPYScore + neutrophil (new) - secondary endpoint
It is possible that the standalone is primary endpoint because it is the root of all the secondary stuff.
If the secondary stuff is well designed-in into the trial from the start, it is part of a whole picture. Which I assume Faulkner and the FDA consultants have vetted to ensure it paints the picture they want to paint.
The secondary endpoints were significant enough that APPY prolonged the trial so they would have sufficient data for the secondary endpoints. This itself MAY suggest pretty good results for the secondary endpoints - since it potentially could be the "below cut-off" group which is already small for combination tests. If it is too small, APPY would need to have more data to give it statistical significance. Thus 800 patient trial was expanded until sufficient patients were added.
If APPY has value in combination with WBC to save costs and reduce radiation risk (and that is a well-integrated part of the trial) WHY would that be toxic ? Even if it is the "secondary endpoint".
You also have to understand the primary and secondary endpoint are not totally different things in this trial - all are essentially linked to the APPYScore standalone - which maybe why it is postured as the "primary endpoint" among the host of endpoints.
You have to remember that just CT alone won't have 98 % NPV. You need to add CT with contrast to get that 98 % NPV.
The great advantage APPY has as a screener is that it only needs to be sure (with high safety i.e. high NPV) on the patients it sends home.
All the rest can be handled by ultrasound, CT scan etc. If it is an "iffy" patient, it just leaves it for ultrasound, CT scan to resolve. So CT scan will STILL be useful (but seeing fewer patients).
However, the outcome is that a sizable chunk of patients are saved the cost of ultrasound and the radiation risk of CT scan.
So the more that APPYScore + WBC can send home at front-desk of ER (safely - i.e. with high NPV) the better "economic" value it has also.
part 3/3 ..
One other new thing to add here - one other way to improve NPV would be to be more liberal about inducting patients into APPYScore trial. This would increase the number of normal folks - which would improve NPV. The more normal folks you have the proportion of normal to appendicitis sent home increases (thus NPV increases) - as explained above.
So new 800 patient trial could use protocols which subtly increase the proportion of normal folks in the data. And this would help (i.e. increase) NPV.
Note that much of the analysis in this thread has been on the APPYScore standalone test. This is because the final PR (510(k) submission to FDA) included only the standalone data. The preliminary data included both standalone and the APPYScore + WBC combination.
I had ignored the combination test because it seemed to be a kludge (was a secondary endpoint of the 600 patient trial). And also because it seemed to work in a SUBSET i.e. the "--" and "++" subset (i.e. where both agree).
And because the standalone test was the primary test, therefore much of the examination focused on that.
However, I would like to ask how folks here view the combination test. While it has good results it seems to work only on a subset (which somehow is not as appealing).
Only reason why APPY could not use the combination data was that it was a secondary endpoint - and the FDA is supposedly averse to "post-hoc" analyses or where tests fail the primary endpoint and then try to hack together some subset that works).
With the combination test now as the primary endpoint, this objection should be removed.
However what is preventing people from accepting the combination test as good enough (apart from that "it relies on another test") i.e. from a data standpoint.
Is the general feeling that the combination test would be a smaller application or "niche" use of the test ?
Would appreciate feedback on this from the physicians on the board as well as others.
Quote: "Only reason why APPY could not use the combination data was that it was a secondary endpoint - and the FDA is supposedly averse to "post-hoc" analyses or where tests fail the primary endpoint and then try to hack together some subset that works).
With the combination test now as the primary endpoint, this objection should be removed."
*i think you got it wrong, the current test clearly states that the primary endpoint is the APPYSCORE standalone test, not a combination test. If the primary endpoint fails (which already failed once in previous test), FDA will not approve the device. There is a reason why it is called a primary endpoint, cause it is the primary results that FDA considers.
Based upon the data set, the interim analysis determined that the trial is adequately sized for the primary endpoint goal, the use of AppyScore test alone. However, the data set gathered for the interim analysis was not sufficient to provide guidance, at this stage, on if the trial is adequately sized for the secondary endpoints -- using AppyScore in combination with either white blood cell count (WBC) or neutrophil count. The Company and all participants in the study remain blinded to the data until the close of the study and there can be no assurance of the outcome of the study based on the interim analysis.
part 2/3 ..
Examining the weaknesses in the 600 patient "failed" trial that APPY earlier identified (conference audio following release of prelim data):
- reducing "mild appendicitis" - one hospital had inordinately high number of "mild appendicitis" cases (which tend to fall below cut-off). In addition the comments suggest this one hospital had a really (actually good for patient) slow process where they were more likely to monitor the patient. It is not clear if APPY also meant that they were testing the same patient more than once - leading to same patient appearing more than once in data - which would tend to increase disease incidence (and harm i.e. lower NPV numbers).
- reducing patient loss to impaired followup. APPY was limited to 3 calls and if patient was unreachable, those patients were EXCLUDED from trial. APPY thought this may have also harmed the data.
Earlier I had suggested that the new APPY management may have taken measures to ensure those "mistakes" don't happen again:
- mild appendicitis - although not suggested by APPY, it is possible that they can choose inner city hospitals i.e. choose locations where patients are reluctant to come to hospital until compelled by severity of disease. Thus you can ensure that you get more "mature" appendicitis cases (fewer "mild appendicitis"). Thus yielding better NPV and sensitivity results.
- better followup - APPY had suggested this was a weakness and they are now choosing university/research hospitals who have better experience of ER trials (presumably this may mean they will be more open to more intrusive followup of patients).
However regarding the impact of this lack of followup (220 patients were excluded - mostly due to inability to followup, while they had a total of 586 patients i.e. "600" patients) - it is not clear if they mean that:
a) patients lost to followup may have been correctly identified by APPYScore as appendicitis, and loss of these harmed results. Presumably they may have gotten admitted to a different hospital (being disgusted at being sent home by first hospital perhaps ?) and been treated there (thus hard to reach at home ?)
b) patients lost to followup may have been mostly normal folks (who were hard to reach at home - may have been at work ?) - thus essentially filtering away the normal folks (which tends to harm i.e. lower NPV).
Earlier I had raised the possibility that (as in (a)) the appendicitis cases maybe fewer in this "lost to followup" group because these are folks who were sent home - so should be generally healthier set. This is probably TRUE.
However I later wondered if it also means that the majority of appendicitis cases would tend to be "mild appendicitis" (which is not good for NPV, sensitivity if these folks are included back into trial).
So on the whole, I am not sure now if improving followup will actually help or hurt.
However reducing mild appendicitis by choosing hospitals astutely WILL surely help both NPV and sensitivity.