Replies to post #705957 on NorthWest Biotherapeutics Inc (NWBO)

[learningcurve2020]

Once the trial effectively became single arm, and with no ECA patient level characteristic similarities, the trial was compromised. Do not discount the screening selection of only the "healthiest" patients in order provide enough time for L to kick in and to monitor for PFS.

To the best of my knowledge regulators never approved the trial changes in real time.

>>It is essential to have constructive input from regulators on the study protocol, analytic approach, selection of RWD sources, endpoint definitions, and other key aspects of the project to avoid miscommunication and surprises down the line.

[dennisdave]

I will assume this is simply ignorance and not baseless pumping,.

Then you should have made more clear in your previous post what you meant (ECA and not DCVAX patient's data). I can't help it if you are just throwing spaghetti around to see what sticks.

The patient level data in question is from the ECA.. NWBO does not posses it, so cannot provide. There is a reason why the FDA and others state this level of data is needed. And this submission does not have. And even the JAMA paper notes it, are they lying?

ofcourse NWBO does not possess ECA data. But again what is your problem? The inclusion and exclusion criteria that were maintained to admit patients to the DCVAXL treatment arm were always spread out to echo the same as nGBM patients in general and FDA-approved. Therefore one can easily scientifically assume that the DCVAXL data can be compared to ECA data because the ECA data not surprisingly echoes the same as general nGBM patient data. This is confirmed with rGBM compared data.

Moreover, Mr ignorant Basher, if NWBO ignored the FDA mandate to (for logical ethical reasons) have the placebo GBM patients NOT randomize to the treatment arm and thus die to be able to compare this data with the treatment arm then that would have made no difference with ECA comparing. You see my ignorant basher friend every person is different. To imply that the medical trial could have only been conducted by comparing a placebo group with a treatment group is BS not even considering the ethical problems this presents. A placebo group (with SOC treatment of course) is by definition a different set of people besides age and gender than a treatment arm simply because as I just said every person and patient is different. Therefore comparing DCVAXL-treated patients with ECA is the same as comparing it if they had compared it with a placebo group if not much better.

Crickets

Are you saying the video posted here was a fake?

Provide me a link to a video where LL implies that she expects DCVAXL not to be approved but she then has confidence that it will continue as RTT.
SHOW ME THAT VIDEO NOW. Yes Im saying alomst everything you say is fake

[sentiment_stocks]

Yes, the external control arms do not have patient level data. But the statistical analysis plan did utilize an adjustment method (the Matching-Adjusted Indirect Comparison - MAIC) to adjust for the lack of this type of data.

Adjustment for Individual Patient Characteristics in ECPs
A fourth set of analyses was conducted to adjust for differences in the individual patient characteristics in the DCVax-L cohort vs. the ECPs. Propensity score matching could not be used because the ECP data were not accessible on a patient-by- patient basis, despite efforts to obtain such data. However, the percentages of specific patient characteristics were available for the ECP. Accordingly, we used Matching-Adjusted Indirect Comparison (MAIC) methodology (widely used in health economic analyses) to adjust for even small differences and re-assess survival outcomes6-8. This methodology applies a weight to each individual patient in the DCVax-L population in such a way that the sum of the weights for patients in each category for a characteristic achieves a match with the external control population. By way of example, if the external population included 50 males and 50 females and the DCVax-L population included 60 males and 40 females, the males would need to be down-weighted to achieve the required 50:50 balance. Applying a weight of 0.67 to each individual male and a weight of 1 to each individual female would achieve that balance, resulting in a new population with effectively 40 males and 40 females. When there are several characteristics to be matched (simultaneously), the mathematics by which the weights are applied to individual patients becomes more complex.

This matching was done on the characteristics of age, sex, race, MGMT methylation status, and KPS score, combined with one of either extent of resection or with residual disease. The MAIC weights required to adjust the DCVax-L cohort to match the characteristics of the external comparators were calculated in the statistical program ‘R’. These weights were checked to ensure that the characteristics of the re-weighted DCVax-L population matched those of the ECP, as well as for any outliers (particularly large weights) that could strongly influence the results of the analysis. Results between the unweighted and weighted (or matched) analyses were compared to confirm the results.

Also, please note... there were three additional analyses used:
1. Matched Comparator Studies for External Comparator Populations (ECP)

The nGBM and rGBM ECPs were determined by an independent expert firm (York Health Economics Consortium), and were comprised of patient data from the control arms of contemporaneous formal RCTs closely matched to the current study based on 14 criteria pre-specified in the SAP (7 for nGBM and 7 for rGBM). For nGBM the criteria included: 1) contemporaneous study time period; 2) reported outcomes (including survival); 3) same standard of care used (radiation and
temozolomide); 4) randomized study design; 5) patients aged >18; 6) KM curves available for survival and for subgroups; and 7) publication in English. The criteria for rGBM at first recurrence were similar., with one difference being that lomustine, bevacizumab or best supportive care were allowed as treatment for the control arms. Based on these criteria, 5 RCTs for nGBM and 10 RCTs for rGBM were matched to our study, and were designated as the ECPs in our SAP (eTable 1). These studies also met the “fit for purpose” criteria outlined by Mishra-Kalyani, et al.1 . The studies that contributed to the ECP provided high quality survival data, with overall lost to follow-up (ltfu) rates of 2% or less (except one study2 that had an ltfu rate of 6%). Similarly, the DCVax-L study had less than 2% of patients lost to follow-up.

2. Validation of the ECP Approach

To further validate the accuracy of the approach and rule out false positive results, each RCT comparator study was evaluated using the same [pooled] ECP that served as the control for the DCVax-L study, instead of using the original control arm in that RCT comparator study, in order to determine whether using this ECP would have changed the outcome originally reported (i.e., primary endpoint met or unmet). This analysis showed that the RCTs that originally failed to show statistically significant survival benefit also failed to show significant survival benefit when the treatment arm of the RCT was compared to our pooled ECP instead of to the original RCT controls, and the one RCT that had a positive outcome was also positive under this ECP approach.

3. Sensitivity Analyses

Although the ECP comparator studies were closely matched with the current study, small non-significant differences could potentially impact outcomes. To address this and further validate the nGBM ECP, we conducted 5 sensitivity analyses, removing each of the 5 comparator studies from the ECP, one at a time. We also conducted a 6th sensitivity analysis, removing 2 of the 5 nGBM comparator trials3, as it was unclear whether they excluded patients with early progression at the time of enrollment, as was done in our study and in 3 of the 5 comparators2,4,5.

https://cdn.jamanetwork.com/ama/content_public/journal/oncology/939074/coi220066supp2_prod_1674054620.80505.pdf?Expires=1723969269&Signature=upcw0yGHuhr3QjYaug7VkxKQ2OCQ2zAuqQJjGgRsCcrew5SKzUHDRRzA6mdDNA6FG3qY8CCazgAdqdFe6BHxuZhQn5QkjC3XBquNybj7HQ7epRJLj5VtvzfsoTLr9OR9iph5dtmCZAPI6QaHI3qYXoZscHX0IHGIiUxzc8Sudfe1p68DCxQbK38865Ajq8cDkmTEPp21zNPJNQt2kChQLI-OtWv0HS4h0u~wjQOLK9SxbqmccmFLeC-OVnjb66DtsClG1ia61BYKicL8cKlfrXSE1kGbPBY8npfDu6xFOW073cVYO-XuVHCQnHT-f6EkVen84zmzUAViPxFQXEHNVQ__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA

All of these additional methods will hopefully prove satisfactory to overcoming any qualms about not having had access to the patient level data that would have had to have been supplied by competing companies (lol).

So there's (all) that, which should help to satisfy the scientific minds in control of approvals, even if it doesn't satisfy those greedy little hedge fund minds.