It appears you have not yet Signed Up with our community. To Sign Up for free, please click here....



Cancer: Prostate Message Board


Cancer: Prostate Board Index


1. [B][U]Highly Skeptical of Chinese Study Conclusion of an Advantage for Surgery Over Radiation For High Gleason/Lower PSA Men: Study Appears Seriously Flawed and Unsupportable to Me, also Contrary to Reliable Research
[/U][/B]

(Part 1 of 2 - Post is in two parts due to length restrictions.)

2. Some of us just want a person exuding authority to take control of our case and tell us what to do. Others of us are searching for evidence to inform our therapy decisions so that we can work in partnership with our doctors, even changing doctors when that seems wise. The following Chinese study was raised in a previous thread as possible evidence that surgery is superior for patients with higher-risk prostate cancer. I reviewed the complete paper, and found the study to be inconsistent with the preponderance of credible evidence I have seen and reviewed, indeed upon which I based my choice of radiation for a once life-threatening case. [I][B]The short version of this post: the research design artificially disadvantages radiation in several ways, yielding an apples-to-grapefruit type comparison, and the unweighted data still, despite several artificial research design disadvantages, show an advantage for radiation.[/B][/I] The following detailed review spotlights typical critical flaws found in this and many other studies where surgery appears to have better results than radiation for higher-risk prostate cancer (with low-risk prostate cancer arguably a draw, but with active surveillance the best choice for most).

I am numbering main paragraphs to foster understanding and discussion.

3. Fortunately, I have an unusually strong background in statistics and experimental design (see post #42, https://www.healthboards.com/boards/cancer-prostate/1048766-club-membership-cards-mailed.html), and that has enabled me to find credibility in the work of the Prostate Cancer Results Study Group (see link later), which essentially demonstrates, with abundant but arguably not yet conclusive evidence, that radiation plus ADT appears to be a substantially better choice for most patients with higher-risk prostate cancer. Therefore, I spent some time analyzing this Chinese study, including trying to communicate with the Chinese research team and succeeding in communicating with one of the doctors who did a pre-publication peer review of the study. This post is my analysis, and the title is my conclusion for those who just want the bottom line.

4. [B]Which Study[/B]: This is the study mentioned in several posts that was initially mentioned in post #1 of this thread at 02-07-2020, 12:15 PM by Djin Tonic (https://www.healthboards.com/boards/cancer-prostate/1048828-advantage-rp-over-rt-subgroups-high-grade-pca.html). The link to the study, entitled “Survival Significance of Patients With Low Prostate-Specific Antigen and High-Grade Prostate Cancer After Radical Prostatectomy, External Beam Radiotherapy, or External Beam Radiotherapy With Brachytherapy” [2019, Full Text] is https://www.frontiersin.org/articles/10.3389/fonc.2019.00638/full. I gave some preliminary comments on the study in post #29, 02-13-2020, 03:00 PM, but at that time I lacked confirmation of several important points, which I now have.

5. [B]Why Focusing on This Study Can Help Us Understand Typical Flaws in Studies that Favor Surgery Over Radiation for High Risk Prostate Cancer Patients, Undermining Their Conclusions[/B]: This study contains several of the common critical flaws in studies favoring surgery, flaws that often produce an “apples-to-grapefruit” comparison that is misleading and not useful rather than a useful “apples-to-apples” comparison. For those who don’t like fruit analogies, let’s try racing – the 100 yard dash: the study is like giving the gold medal to a racer who gets to start a second ahead of time against an opponent who has to start a second later and carry a 20 pound extra weight, with the finish based on the chest passing 100 yards for the first runner versus the back heel passing for the second. A lot of studies have these critical design flaws. Understanding these critical flaws can help patients sort the wheat from the chaff in other studies.

6. [B]Some of the typical flaws, occurring in this study[/B] in my opinion are: failure to adequately account for differences in average ages and other current serious health conditions (“comorbidity”) in men in surgery groups versus radiation groups; failure to adjust for the benefit of adjuvant or salvage radiation in men in surgery groups; failure to treat Gleason scoring equally in the surgery and radiation groups; and failure to treat staging information equally in the surgery and radiation groups. Such flaws can and fairly often do completely undermine the validity and conclusions of published medical research studies, in my opinion as a now savvy layman (but no enrolled medical education). Awareness of such flaws helped me sort through competing claims of research studies when I was choosing whether to take a curative shot with surgery or radiation, which I resolved strongly in favor of radiation (2013) for my high-risk case. The criteria used by the Prostate Cancer Results Study Group in selecting studies deemed credible as evidence go a long way toward minimizing these flaws, and are stated at the group’s website, http://www.prostatecancertreatmentcenter.com/prostate-cancer/study-group/ .

7. [B]The Prostate Cancer Results Study Group (PCRSG) results[/B], which group non-recurrence results from many qualifying studies for different treatments and show them graphically based on risk group and length of follow-up, helped me. The PCRSG used several techniques to avoid or minimize the flaws evident in this study. Here is a link to their work, based on a 2012 journal publication: http://www.prostatecancertreatmentcenter.com/prostate-cancer/study-group/ Generally, with research on low-risk patients being of limited value since the advent of active surveillance as an option, their results for intermediate-risk and high-risk patients show superior results for radiation, often with androgen deprivation therapy (ADT). It is also clear that surgery is successful for fairly substantial percentages of men with higher-risk prostate cancer, just not typically nearly as successful as radiation. The Study Group’s graphs for “higher risk” men show surgical success rates of up to around 60%, long-term, in some studies, though substantially lower in other studies, with the bulk of radiation studies showing substantially superior results. Different types of radiation and combinations, as well as other therapies, are displayed using symbols of different shape and color. Patients who really want to get into the grass roots can go to the references for the study, a list that shows each study that was a data point in the results graphs. One shortcoming of the study is that it was based on data going into the 2012 paper. That meant that many of the radiation patients in the long-term results were not treated with advanced imaging guidance, and may not have been treated with at least 18 months of ADT, which is now known to be important for achieving best results for high-risk men (with a short course best for intermediate risk men). Also, the minimum EBRT radiation dose for inclusion in the PCRSG study was 72 Gy, which is now known to be inferior to a dose of 78-81 Gy. In short, results for patients being treated in more recent years should be even better.

8. [B]My Review and a Supportive Confirmation[/B]: I have now had time to review the study carefully and get some semi-confirming information regarding its data on Gleason scoring and staging. Two of the Chinese researchers listed as corresponding authors have not responded to my inquiries, stated in my post #32 of 02-20-2020, 03:50 PM, but one of the reviewers did respond that my first two assumptions, in those inquiries, appeared to be correct, adding that SEER data did cover radiation but with uncertainty about the timing. The following is based on assuming that the Gleason scoring was based on the higher of the original biopsy OR the post-prostatectomy specimen, which is almost but not quite stated in the paper, and that staging data were also based on the higher of pre-RP clinical data or post-RP pathology and other data, whichever were higher, which is not stated in the paper but which appears to be the case based on data in the paper and the peer reviewer’s statement that that appeared to be the case. As noted earlier, I have had no enrolled medical education, but I had an extraordinarily strong statistical and experimental design background in academic study, more than 225 classroom hours, plus strong life experience in analysis. In contrast, my impression is that most pre-med and MD students, other than those with additional research-credential study, likely have only one course in elementary statistics under their belts when they start practicing, equivalent to about 45 classroom hours. I have been following prostate cancer research for more than 20 years, since my diagnosis.

9. [B]The Study’s Conclusion, and My Impression[/B]: The study concluded that: - “RP patients with low PSA levels and high GS [Gleason scores] had better OS [Overall Survival] compared to either EBRT [External Beam Radiation] or EBRT+BT [Brachytherapy], while RP and EBRT+BT resulted in significantly lower PCSM [Prostate Cancer Specific Mortality], compared to EBRT.” – appears seriously misleading to me as the study was an apples-to-grapefruit comparison. While the second line of the conclusion is more in line with what I would expect – “Moreover, EBRT+BT and RP were associated with similar survival of patients with age of > 70 years old, or PSA levels of ≤ 2.5 ng/ml.” -, even there elimination of the flaws, were it possible, would show superiority of radiation, in my opinion.

10. [B]Key Flaw in the Study: Unequal Gleason Scoring[/B] The study used the higher of two possible Gleason scores for each patient in the study: “GS provided by the SEER program represents the highest GS found during a surgical or non-surgical biopsy.” This means that the higher Gleason score from either the clinical, pre-RP biopsy or the pathology biopsy of the removed prostate was used. As we patients are well aware, radiation patients do not have the latter biopsy. Therefore, in determining which patients were “high-grade” in this study, we KNOW that the radiation patients were all high grade based on their original clinical biopsy, but it is likely that a portion of the surgery patients had Gleason score 7 or even Gleason score 6 or lower clinical biopsy results that were only later upgraded. Given that the surgery patients had PSAs of 10 or lower, and that at least (see below) two thirds of them had stage T1 or T2 cancer per Table 1 in the study, that would have made them “Intermediate-risk” or “low-risk” patients based on the initial biopsy. It is known that pools of patients considered low- and intermediate-risk have considerably better outcomes of patients considered high-risk, so this establishes that you have a group (actually two groups) of high-risk radiation patients (EBRT alone, and EBRT plus seeds), competing for survival against a mixed group of low-risk, intermediate-risk and high-risk surgery patients. So even though the RP group was determined to be high-risk, , many based on the post-surgical pathology, that RP group was bound to be a more favorable group of high-risk patients, such as less extensive spread within the prostate, and likely fewer high-risk individual tumors in each patient’s prostate, than in the two radiation groups. As we know, biopsies sample only a small proportion of a patient’s prostate, whereas a post-surgery biopsy is comprehensive.

11. [B]Key Flaw in the Study: Unequal Staging Data[/B] Similarly, it appears that the T1, T2, T3, T4 staging information from each patient was based only on the initial, clinical biopsy and other (DRE, etc.) staging data for the radiation patients but was based on the higher of that data or post-surgery data (such as positive margins, extracapsular extension, positive seminal vesicles, etc.) data for the surgery patients. The reviewer concurred with me that this was probably the case, but evidence in the study also suggests it is the case. Consider the proportions of men with each stage in Table 1 of the study:

[B]Table A. Percentages of patients in the Chinese study with each stage of prostate cancer[/B]

Stage……………………….RP…………………EBRT……………EBRT+BT

T1…………………………. 0.5% …………… 54.5% ............. 60.6%

T2……………………….. 62.5% …………… 39.2% ............. 33.7%

T3……………………….. 34.7% …………….. 5.5% ............... 5.5%

T4…………………………. 2.4% …………….. 0.8% ............... 0.2%

Totals…………….…. 100.1% …………. 100.0% ………….. 100.0%

12. The line for Stage T1 patients is our first clue. As all of these patients had a PSA score of less than or equal to (<=) 10, indeed with the average for each group being around 6, it is likely that many would have been first alerted by a rising and/or elevated PSA score and a normal DRE (Digital Rectal Exam). Indeed, that is what we see for the two groups of radiation patients, where more than half had negative DREs as evidenced by the designation of stage as T1, meaning no nodule (and nothing visible by imaging if done). Surprisingly, the RP group had almost no patients staged as T1; this suggests that post-surgery results upgraded the initial staging, and that probably accounts for the proportionately much higher percentages for stages T2 and T3 for RP patients than for either group of radiation patients.

13. Moreover, in the US it has been unusual to treat prostate cancer patients with radical prostatectomy if the disease had already escaped the capsule, as is indicated by T3 staging. Yet, as we see in the chart, just over a third of the RP patients were stage 3. This too suggests it is likely that RP patients were upstaged as a result of post-prostatectomy findings.

14. The significance of this is that RP patients in the study are considered to be at much higher risk, based on staging, than their radiation counterparts, though this is probably false as it is based on a second cut at the staging apple provided by post-surgery observations. In other words, if pathologists had somehow done biopsies on prostates removed from all the radiation patients, they probably would have upstaged them to about the proportions seen for the RP patients, or, as the RT patients were older, somewhat higher in average staging; of course in reality, that cannot be done. However, when making treatment decisions, many of these RP patients and their doctors likely thought they were low- or intermediate-risk patients based on staging alone: likely low-risk, clinical pre-prostatectomy stage of T1 or T2 based on the analysis here, low or intermediate risk Gleason prior to upgrading after surgery as discussed above, and a PSA not exceeding 10 as a requirement for inclusion in the study). So again, this suggests that the truly and originally high-risk radiation groups are being compared to what would have been, based on pre-operative evidence, an RP group with a mix of low-, intermediate- and high-risk patients. That kind of situation pretty much guarantees artificially better results for the RP group. It’s an apples to grapefruit comparison from the get go.

15. [B]Key Flaw in the Study: Overall Survival Results - Unequal Starting (At Treatment) Average Ages in Each Group, with Aggravating Elderly Factor Confounds Claimed Superiority of RP for Overall Survival[/B]

Not only were the radiation patients on average significantly older than the RP group, but a substantial proportion of them were entering the age range where many men near the end of their life spans, which, though I don’t feel that elderly at 76, a lot of people would call elderly. For elderly men, who are losing friends and peers to death as I am, upcoming life span is typically not long, a matter of a handful of years, especially for those of us with major challenges to our health; in contrast, men in their 60s (the RP group average) have many years of life before them, and typically have fewer threats to their health. This obviously is going to affect the “OS” (Overall Survival) result of a study by itself, regardless of disease or treatment. Again, this means the comparison in this study is apples-to-grapefruit and not apples-to-apples.

Table B. Here are the [B]age differences[/B]:



Age/Life expect. …….RP…………………EBRT……………EBRT+BT

Average (median)……64 ………………. 71................. 68

Avg. birth year …… 1956 …………… 1949 ……………..1952

Avg. added years* … 19 …………………. 14 …………….. 17 (Keep in mind, this is US population wide,
……………………………………………………………………………………..not just patients with a major disease like PC.
………………………………………………………………………………………Insurers are not fond of issuing
………………………………………………………………………………………life insurance policies to us.)

Avg. total years
At death: ………………….. 84 …………………. 86 ……………… 85

* Additional years of expected life for the average male who has attained the median age of one of the groups in the study, that is 64, 71, or 68; source: https://www.ssa.gov/OACT/population/longevity.html

Continued in Part 2
[B]Part 2 of 2 re Chinese Study discussed in Post #2[/B]

16. Now let’s take a look at what the average ages would be for men in each group in the study at the start, after 3 years, after 5 years, and after 10 years.

[B]Table C. Average Ages of Men in Each Group Projected at Treatment to 3, 5 and 10 Years Since Treatment[/B]

[B]Age/Life expect[/B]. …….[B]RP[/B]…………………[B]EBRT[/B]……………[B]EBRT+BT[/B]

Average (median)……64 ………………. 71................. 68

+ Three Years ………… 67………………… 74 ……………….. 71

+ Five Years …………… 69 ………….……. 76 ……………….. 73

+ Ten Years …………... 74 ……………….. 81 ………………… 78

Avg. total years
At death: ………………….. 84 …………………. 86 ……………… 85

17. What this shows is that the EBRT group is not only 7 years older than the RP group (and 3 years older than the EBRT+BT group), but it is already, at treatment, in the age range where we tend to lose peers to death, in sharp contrast to the RP group, where death of peers is rather unusual. By five years since treatment the EBRT group is definitely in the “elderly” range, while the RP group is not. In other words, the odds of death from other causes than prostate cancer are substantially higher for the EBRT group when it hits the five year mark, and especially when it hits the 10 year mark. The final line of the table shows that the EBRT group, at 10 years from treatment, would be just 5 years from expected death for the average more-or-less healthy US man, not one with a major disease like prostate cancer. It makes sense that overall survival at the 10 year point would fall off sharply for the EBRT group from age alone, aside from any treatment effect. To a lesser extent, because the EBRT+BT group is younger, this also applies to the EBRT+BT group. The part of Table 2 in the study for Overall Survival Using Kaplan Meier Analysis is consistent with this analysis, as follows:

[B]Table D. Percentage Overall Survival of Men in Each Group Projected at Treatment to 3, 5 and 10 Years Since Treatment[/B]

[B]Treatment group[/B]. ….[B]RP[/B]…………………[B]EBRT[/B]……………[B]EBRT+BT[/B]

+ Three Years ….. 98.3%……………… 94.0% …………….. 96.8%

+ Five Years ..…… 96.2% ……….……. 86.3% …………….. 92.5%

+ Ten Years ……... 73.5% …………….. 54.7% ……………… 66.5%

18. Table 2 also contains a weighted adjustment section where age is one of the factors taken into account, though the method is not clear. As expected from the analysis above, the 10 year survival for the EBRT group imrpoves by nearly 4 years while that for the RP group drops by 6 years, with a drop of 5 years for the EBRT+BT group.
19. My conclusion is that the overall survival data are not helpful because of the artificial aspects discussed above, chiefly the fact that those men in the EBRT group are, on average, approaching the natural limit on their life spans.

20. [B]Other Flaws in the Study
[/B]
1. It is likely that [B]radiation received by EBRT and EBRT+BT patients was substantially inferior to what would be used today[/B], leading to inferior results for the radiation groups. We do not know the quality of radiation used, and this is critical because old style radiation was definitely inferior in controlling cancer and avoiding a high level of side effects. We do know that the earliest radiation patients were treated not earlier than 2004, the earliest year for inclusion in the study. At that time many US patients were being treated with EBRT doses of 70 Gy or lower, and we now know that a dose range of 78-81 Gy is substantially superior for most patients. Moreover, advanced imaging to aid radiation targeting, which improved effectiveness and reduced side effects, was not widely accepted until at least 2007, well after the start of the study. This applied to brachytherapy as well as to EBRT. Finally, it was not known in the early 2000s that a course of ADT of at least 18 months substantially boosted radiation effectiveness for high-risk patients; unfortunately, the study provides no data on ADT use to support radiation.
2. [B]Adjuvant and salvage radiation data are not provided for the RP group.[/B]. It is a sound assumption that many of the RP patients would have had adjuvant or salvage radiation for their high-risk cases, but the study provides no data on this. In other words, success for the RP group is no doubt improved by radiation, but we do not know by how much.
3. Similarly, [B]use of ADT by RP patients[/B] after surgery is not provided, and this too would have improved results for the RP group.


21. [B]Assessment of the Value of this Study
[/B]
1. [B]Understanding the effectiveness of RP versus radiation[/B]: The study does not help due to the critical and other flaws discussed above.
2. [B]Understanding prognosis for high-risk patients considering surgery[/B]:
The study does not help because high-risk characteristics, such as Gleason and stage, are apparently based on post-surgery findings rather than initial biopsy findings.
3. [B]Understanding prognosis for surgery patients based on post-surgery findings (Gleason, stage)[/B]: the study provides some useful information, especially for younger men like those in the study, with an average age of about 64.

[COLOR="RoyalBlue"]22. I am adding (3/11/2020) some text in blue from my post #29 02-13-2020, 12:00 PM, on DjinTonic's early thread (https://www.healthboards.com/boards/cancer-prostate/1048828-advantage-rp-over-rt-subgroups-high-grade-pca-2.html).

However, some of the facts in the study support the opposite conclusion than the one in the study, which favored surgery; these facts support an apparent superiority of radiation for such high-Gleason score/PSA <=10 patients, and there are serious questions about the approach taken by the urologist authors in trying to support a conclusion that surgery is the better approach. I hope to have a lot more to say about that when I complete my look at the study.

For instance, consider the proportions of patients who were alive in their early to mid-70s, specifically at ages 73 for the RP group, 75 for the EBRT group, and 72 for the EBRT+Brachytherapy groups. Remember that these were all patients in the SEER database that covers about 28% of the US population, therefore including all from the US SEER areas, who had a Gleason score of 8-10, a PSA of up to 10, and no detectable nodal or distant metastases (N0, M0), between 2004 and 2015.

...........................Survival at.............Overall
Treatment ..........Average Age:.......... Survival


RP............................73........ ........74% (73.5%)


EBRT........................75.......... ......86% (86.3%)


EBRT+Brachy............72............... .93% (92.5%)


This table is from Table 2 in the report, with selection to focus on outcomes at nearly equivalent ages. Now there is a twist here, and the authors could cry foul, but it is still a factual and meaningful view of the data, though ignoring one key fact, which I will put below my signature to allow a little suspense and an opportunity for Board participants to try to figure this out as in a "Who Done It". The table shows a clear pattern of superior survival of radiation patients versus RP patients at approximately equivalent ages.


Now here is data taken straight from Table 2 of the study about Prostate Cancer Specific Mortality (PCSM), and I’m going to use the figures for 3 and 5 years from diagnosis but not for 10 years, because the confidence intervals in the 10 year column are so wide, indicating low-confidence as to the true value of this projection and that the data in the study must be all over the map, not in a clear, fairly tight pattern:


Prostate Cancer Specific Mortality (PCSM – Dying Due to Prostate Cancer, rounded and exact, the lower the percentage the better)


...............................PCSM at................PCSM at
Treatment................3 Years:..................5 Years


RP.........................6% (6%)#............16% (16%)*


EBRT.....................2% (1.9%)#...........5% (5.3%)#


EBRT+Brachy.........1% (1%)#..............3% (2.6%)#


#For each of these values the confidence interval is fairly tight, indicating that the true value of the projection is very likely very near the indicated value.


* For RP, the confidence interval at 5 years is wide, from a true value ranging from 12% to 22%, with 16% projected as the most likely, indicating that data points are widely spread. The CI interval means that the true value could be as low as 12% mortality or as high as 22%.



It appears that at both the 3 and 5 year points, both forms of radiation have sharply lower death rates, which is counter to the study’s conclusion that RP patients do better. More on that, more reasons why it is probably an unsound conclusion, to follow, time permitting.


(At ten years, RP shows the lowest mortality, but the confidence intervals for RP, EBRT, and EBRT plus brachytherapy are all very wide, indicating little reliability in the projected averages due to data points being very widely spread from each other. Various factors come into play, such as the increasing age of patients, which causes more deaths from other causes and a decreased ability to see what ages would have been like if death came from prostate cancer; increasing age typically means a substantial decrease in the number of data points upon which to base a projection for death specifically from prostate cancer.)[/COLOR]

[COLOR="Green"]My other posts on Djin's earlier thread (https://www.healthboards.com/boards/cancer-prostate/1048828-advantage-rp-over-rt-subgroups-high-grade-pca-2.html), some of which detailed additional problems, were:

#13 02-08-2020, 05:09 PM Not Impressed with the Tilki 2019 study: Surgery vs Radiotherapy in the Management of Biopsy Gleason Score 9-10 Prostate Cancer and the Risk of Mortality

#25 02-10-2020, 01:30 PM A Bit More on the Tilki Study, and What "Immediately Following" Means for Adjuvant Radiation Following Surgery

#27 02-10-2020, 02:24 PMWhack-a-Mole Deja vu All Over Again - (I suspect)

#29 02-13-2020, 12:00 PM Midway Point in Analysis of Chinese Study of US Patients with Gleason 8-10 and PSA of Up To 10 - Radiation Looking Better[/COLOR]

Jim

- - - - - - - - - - - - - - - - - - - - - - - -
Diagnosis Dec 1999 PSA 113.6 (first ever), age 56
Gleason 4+3=7 (J. Epstein, JHU), all cores +, most 100%; "rock hard" prostate with ECE - stage 3, PNI, PSADT determined later 3-4 months; technetium bone scan and CT scan negative; prognosis 5 years.
Later ProstaScint scan negative except for one suspicious small area in an unlikely location. ADT Lupron as first therapy, in Dec 1999, then + Casodex in March 2000, then + Proscar and Fosamax in Sep 2000. Rejected for surgery January 2000; offered radiation but told success odds were low; switched to ADT only vice radiation in May 2000, betting on holding the fort for improved technology; PSA gradual decline to <0.01 May 2002. Commenced intermittent ADT3 (IADT3) with first vacation from Lupron & Casodex. Negative advanced scans in 2011 (NaF18 PET/CT for bone) and 2012 (Feraheme USPIO for nodes and soft tissue). With improved technology, tried TomoTherapy RT, 39 sessions, in early 2013, plus ADT 3 in support for 18 months (fourth round of IADT3), ended April 2014. Continuing with Avodart as anti-recurrence shield. Current PSA remarkably low at <0.01; apparently cured.. Supportive diet/nutrition, exercise, supportive medications during this journey, as well as switches in antiandrogen, 5-ARI, and bone drugs.





All times are GMT -7. The time now is 07:54 PM.





© 2020 MH Sub I, LLC dba Internet Brands. All rights reserved.
Do not copy or redistribute in any form!