In recent years we are faced more and more frequently with patients with middle- and lower third rectal cancer responding completely (ypT0N0M0) to neoadjuvant CRT. And the best treatment strategy for those patients is still evolving. Based on recent evidence and your center experience how should be managed these patients?
Which of the following options is you preferred one and why:
1. Standard rectal resection with total mesorectal excision (open/laparoscopic).
2. Full-thickness excision only (by transanal endoscopic microsurgery-TEM).
3. Full-thickness excision plus endoscopic mesorectal excision(transanal/transperineal)
4. Biopsy only to confirm complete response.
5. Other.
Standard rectal excision,based on inicial clinical staging. Unless if it is an older person with low anal pressures in anal manometry or with severe co morbilities when we consider trans anal excision ( if scar visible) or biopsy to confirm response
Thanks - we have been following up patients with a complete response who we have not operated on. We published our early results in Colorectal Disease, under Dalton et al. We are continually following these patients and looking for factors associated with complete clinical response. The PubMed reference is; http://www.ncbi.nlm.nih.gov/pubmed/21831177
It’s a very interesting question, I read this new paper of the Memorial Sloan-Kettering Cancer Center, http://www.ncbi.nlm.nih.gov/pubmed/23154394 , i believe that a non operative approach could became a first choose for all patient with complete clinical response after neoadjuvant therapy. Actually in our center in Rome, we are following a 35 years-old man with complete response of very low rectal cancer making excisional biopsy every 6 month. After 2 years the patient is alive free of disease. Alternative was a Miles surgical procedure with a reduced quality of life.
We have been been following a group of patients like this for some years - we call it 'active surveillance' rather than 'watchful waiting', regular clinical assessment and serial MRI scans. Some do a TME excision of the scar but we have not done this. Anterior resection continues to have significant morbidity and it never feels right when the path comes back as complete response. I think this situation will face us more and more in the future and maybe one day low rectal cancer will stop being a surgical problem!
Absolutely agree with you Mr. Motson! And we are on the way. The evolution of treatment strategy for low rectal cancer is very similar of that for breast cancer – extended surgery at the beginning, after that - neoadjuvant treatment followed by standard surgery, and more recently – neoadjuvant CRT and organ-sparing strategies. And the key for optimal results seems to be a stratification of the patient. When you look at different TNM stages and risk-stratification in two groups: low-risk (sm1-2, G1-2, L0, V0, tumor3cm), the current evidence is:
1. Grade A evidence – T1N0M0 low-risk pts can be treated with transanal endoscopic microsurgery (TEM) with equivalent to radical surgery results.
2. Grade B evidence – T2N0M0 low-risk pts can be treated with neoadjuvant CRT followed by transanal endoscopic microsurgery (TEM) with equivalent to radical surgery results.
3. Grade C evidence – after neoadjuvant CRT patients staged as ypT0-1N0M0 can be treated with transanal endoscopic microsurgery (TEM) with equivalent to radical surgery results.
What to say – it seems that local excision after complete response is important for proper stratification, and can be the only surgical intervention in most cases. On the other hand there is a need to change protocols for neoadjuvant CRT as in many countries (if not all) early stages (T1-2,N0) are not indicated today for neoadjuvant CRT. And the latter have potential to be the primary curative treatment. As the question with lymph-node involvement remains open with this approach, and especially for high-risk patient we need a solution there. And the isolated endoscopic mesorectal excision seems very interesting (just like lymphadenectomy and organ-sparing mastectomy). There are many ongoing trials and I hope we are close to some answers.
A Plea for Caution
The nonoperative approach, known as “watch and wait” (or "wait and see") has been long associated with the team of Angelita Habr-Gama and colleagues [Br J Surg. 2009, and others] at the University of São Paulo School of Medicine, although now several other teams are pursuing the approach, such as Phillip Gaty and colleagues at Memorial Sloan-Kettering Cancer Center, and Geerard Beets and Monique Maas and colleagues in the Netherlands, among others including Ian Daniels team at Royal Devon & Exeter Hospital (see Ian Daniel's contribution, above). Some considerable long-term results have been reported, although to date the approach remains controversial and we must await mature results from randomized controlled trials before considering this a validated option.
Thus, there are unresolved concerns stemming from the inaccuracies of post-treatment clinical staging, sometimes due to excessively short (~6 - 8 weeks) waiting periods where it is speculated that such shorter intervals may actually interrupt tumor necrosis that is in progress and so prematurely detect transient residual disease, and with wide inter-observer variability as to tumor complete response as found by Spiros Hiotis and colleagues at Memorial Sloan-Kettering.
In this connection, a retrospective review conducted by these same investigators at Memorial Sloan-Kettering, [J Am Coll Surg. 2002] of the clinical and pathologic characteristics of 488 patients from a prospective colorectal database, found that although clinical complete response was a significant predictive factor for pathologic complete response, nonetheless 75% of clinical complete responders had persistent foci of tumor undetectable on preoperative examination or proctoscopy. This suggests that clinical complete response to preoperative therapy is not an accurate predictor of pathologic complete response, with a significant percentage of clinical complete responders having persistent deep tumors or nodal involvement.
In addition, despite largely retrospective study support (and largely from a single series) with only a small Dutch study providing limited positive prospective data, all studies of the association of clinical outcome endpoints and the "watch and wait" strategy exhibited significant heterogeneity in study design and inclusion criteria, staging, and the rigor and duration of follow-up after chemoradiation therapy. Furthermore there was manifest inconsistencies in the definition of cCR, with only partial and erratic concordance with pCR. A critical appraisal and review of this literature to date still finds the weight of the cumulative evidence to support the conclusion that patients selected for observation, but who subsequently do not sustain a cCR may fare worse in clinical outcome than patients undergoing immediate tumor resection. Nor should we be tempted warranted to extrapolate without further assurance than we have to date, from proof of principle in small low - and easily clinically assessed - rectal cancers to more advanced rectal cancers with relatively common nodal involvement. For therapeutic decisions in this context to be sufficiently decisive across a wide spectrum of potential case scenarios, we still require (1) robust long-term prospective observational studies with (2) long duration of follow-up, (3) tightly uniform inclusion criteria, and (4) consistent application of the definition of cCR with demonstrably close concordance with pCR. Only then can we be maximally confidant about the potential oncologic benefit of resection despite attainment of pCR (except of course in patients either demonstrably unfit for radical surgery or refusing surgical intervention).
Constantine Kaniklidis
European Association for Cancer Research (EACR)
All these concerns are well known and any physician dealing with patients with low rectal cancer kept in mind them. However may I clarify again that my question is related to patients with pathological complete response. And at our center this means that he/she underwent transanal endoscopic full-thickness excision of the scar to confirm CR.
The answer is surgical treatment. W´ve a series of patients with ypT0, but lymph node involvment. These lymph nodes are small and cannot detect neither with MRI nor multislice CT or endosonographic examination. If you leave tumor in situ, the procedure is without curative intent. As long as lymph node involvment cannot be detectable accurately, the therapy of choice ist surgery.
There is no prospective study on watch and await approaches in pCR after neoadjuvant in rectal cancer. It is just if the patient is unwilling to undergo surgery or if clinical condition contraindicate surgery that it shouldn't be performed. Moreover, how would you evaluate CR if you do not see the surgical specimen? RMN and EUS are clearly suboptimal in determining differences between isolated tumor remnants and sc tissue. We have seen ever so often island of surviving tumor cells in the context of what seemed to be a CR. The question should rather be whether to complete neoadjuvant treatment (we do 4 week capecitabine 825 mg/sqm bid + RT) with another 6-7 course of 5-FU based treatment. Again, this depends on initial radiological assessment, on surgical pathologic evaluation, and on patient's conditions. Again, there is no large prospective randomized, control-treatment study to show whether pCR without terminating the full scheduled regimen in non-inferior in terms of relapse compared to completed the aformentioned regimen...
Alexander: Thanks for the clarification of your center's policy of confirmation of CR via TEM full-thickness excision, not information originally available in your initial query, but I appreciate your point and we are on that score in agreement. As to how well-known the issues I raised are, is, however, somewhat arguable I would respectfully suggest given that discussion and debate on the matter continues to appear regularly in the literature (and including in some panel discussions at some conferences I attend, including NCCN), and as an internal reviewer for several journals in the field, I've encountered the issue, pro and con, in several articles just within the last 18 months, and with several exchanges on all sides of the issue breaking out sporadically more than once recently (for example, Angelita Habr-Gama and Rodrigo Perez at University of Sao Paulo, Brad Champagne at Case Western Reserve Medical, Center, Richard Goldberg at Ohio State, Tim Nguyen at the Cleveland Clinic, and Theodore Saclarides at Loyola University all debating the matter within the pages of Gastroenterology & Endoscopy News just this past December). From this, my sense is that we still have not reach deep consensus, so it may still be valuable to put the concerns out there.
Zoppoli: Point taken - my reference was to the Dutch prospective study of Geerard Beets and colleagues at Maastricht University of cCR responders post-chemoradiation neoadjuvant therapy, compared with a small control group of patients with pCR after CRT and TME, but my intent was to expand the discussion into the wider focus on wait and see strategies in general, not just vis a vis pCR.
Zoppoli: Was your own contribution truncated prematurely? (something that I observe sometimes happens upon editing).
Constantine: it was a "..." open to further discussion :-) Great one indeed, it's been transformed into a consensus conference blog!!!
I apologize for my naive question, since I am not an oncologist. Do some biomarkers, gene and/or protein signatures or whatever exist that can guide towards a proper (also from an ethical point of view) decision?
Nicola: What is known basically boils down to few small (i.e. < 40-60 pts) studies assessing microarray gene epression profiles or miRNA microarray profiles. Nothing big enough to generate a real classifier in the neoadjuvant setting. We're actually collecting frozen specimens from pre-treatment samples, but it's not as easy as one may think because of the easy degradation of fresh rectal biopsies and unavoidable contamination by bacterial RNAses and local anesthetics. Thus fur, we're at number 41. Also, there's a good paper (I think it's in Nature last year, but I may get wrong), showing that no biological differences fundamentally exist between colon and rectal cancer. The issue is that one should find biomarkers to predict radiation treatment combined with 5-FU based chemo, and that this implies generating ad hoc studies for this particular type of treatment. People have done it. As I said, studies are too small to be anything more than explorative in nature, and have not lead to changes in clinical practice yet.
Nikola: to the best of my knowledge - still nothing conclusive and widely useful. Several groups are working on the subject. And the questions are more than answers. The pathologists from our team recently noted (paper in press) that the density of the tumor-infiltrating CD68+ macrophages correlates inversely with TGF-β1 expression as well as with several histopathologic tumor characteristics. They studied 210 unselected pts with CRC and low CD68+ infiltration was found to be unfavorable independent prognostic factor for both overall and disease-free survival. As not only the tumor characteristics but also the host response is always important we try to keep in mind such data. But the big questions raised by anecdotal experience from several teams still wait the answers. All we had some 15-20 or more long-term disease-free survivors treated with RT/CRT ± local excision. And those pts usually are unfit for- or refused major surgery. As we seen that RT/CRT ± local excision can cure some pts we would be happy to know how to expand safely this approach without compromising oncological outcome. And interestingly the groups of patients that looks most curable (as T1-2N0M0) today are not indicated for neoadjuvant CRT, but that is separate discussion.
Actually it is not possible to know if a Patient is N0. So who would take the risk to leave lymphnodes with tumor if the patient are operable...
Dear Marc-Olivier,
Actually some people do that in selected group of patients. And the published results from their randomized studies does not shown difference in outcomes:
http://www.ncbi.nlm.nih.gov/pubmed/22864880
http://www.ncbi.nlm.nih.gov/pubmed/8797643
It is clear that in some patients major surgery can be replaced by organ-sparing intervention. It is also clear that safe N-staging is still problematic, as we are looking usually for size of lymph nodes which is not the best predictor for presence of metastatic involvement - approximately half of the metastatic nodes are smaller than 6mm in many studies. All we need more accurate preoperative N-staging tools, but it seems possible in the meantime to do something better even for highly selected group of patients. However I myself still do not feel comfortable to propose local treatment for a fit for standard surgery patient, but I see that frequently we are doing too much in cases where pathologists cannot find anything malignant in the specimen after CRT.
Thanks Dr Ofner - the difficulty is that we are assuming that the biology of these tumours is the same in the primary tumour and nodes in response to the treatment and we are assuming that there is no morbidity nor mortality risk from the treatments that we give, hence the answer cannot be a simple - operate on all. Long term data on yT0yN0 tumours has a low rate of recurrence, and yet there is both recognised mortality and mortality from the treatment given. We need more information on this and the answer is not through a randomised trial but using statistical process control and multinational cooperation as the numbers of these cases are small and the units involved need to work to the same standards.
Surgery is the treatment strategy to be followed in these patients, taking into consideration the major evidence based scenario available till now.May be it changes in future........?
Dr. Torday, here is the link:
http://www.cancerworld.com/Articles/Issues_52/Cover_Story/Vincenzo_Valentini%3A_putting_knowledge_to_work.html
May be the recently proposed PROGRESS (Peri Rectal Oncologic Gateway for Retroperitoneal EndoScopic Surgery) procedure could be an answer to my question. I would like to hear your comments about it - video is available at the WebSurg:
http://www.websurg.com/MEDIA/?doi=vd01en3937
These issues will be adressed in detailed by some of Europes finest at the 4th Biannual STERC conference in Bergen, Norway 1-3 may 2013. Check out http://www.sterc.no
to see the program. The sessions includes watchful waiting series from Mastricht (G.Beets) and Oxford (G.Brown), the duch CARTS trial and general update from R.Glynne-Jones.
Jo:
Thanks Jo for that alert re the STERC 2013 Conference: I anticipate that we may be hearing the first preliminary findings from the Geerard Beets Watchful Waiting Trial (aka, Minimal Invasive Strategies for Good and Complete Response to Chemoradiation in Rectal Cancer [NCT00939666] since I understand from the principals (personal communication) that the final data collection has completed and quantifications of the primary and secondary outcome measures (LR, OS, DMFS) are even now being finalized, a good year in advance of the Danish Vejle Hospital Watchful Waiting Trial (NCT00952926) is slated to completed final data collection (LQ, 2014). And Rob Glylnne-Jones will I am confidant underline the concerns I raised above, further reinforcing the absence of any reliable modality to date that can evaluate a clinical complete response that in any clinically relevant way correlates with pathologic complete response. And Chris Cunningham and Göran Kurlberg promise, and are sure to deliver, a closely reasoned best-of-evidence consensus statement which is to be integrated with several EU-based guidelines that are pending on the issue.
According to our experience the results are excellent if the surgical treatment will be based on the pre-neoadjuvant CRT stage. Rectal Ca of middle and lower third with complete response could be treated by TEM resection to avoid the stoma, but rectal Ca of upper third must be treated by low anterior resection.
I agree with some of the points above regarding caution. We found in 2 large centres a 10% complete response ( Neoadjuvant chemo-radiotherapy and rectal cancer: can the UK watch and wait with Brazil?S M Nyasavajjala, A G Shaw, A Q Khan, S R Brown, J N Lund Colorectal Disease 10/2009; 12(1):33-6.) but even though some specimens had complete mucosal response, there remained viable cancer in lymph nodes and in layers beneath the mucosa. We would still advocate removal of all rectal cancers, as at the moment you just cannot tell. If the patient feels very strongly that they do not want an operation after CRT and subsequent good response on MR then that is their decision to make after being informed of all the pros and cons.
The results of "Wait and see" of Habr Gama are very interesting but the main problem is that these results exist only in one team: Habr Gama's team. In the others team the results are not so good. We use this attitude only in patient with very high risk of morbidity or with a short survey . In the most of case, a complete clinical response has to be operated by total mesorectal excision
question is how accurately can a complete response on imaging predict a complete pathological response... and whether there was lymph node disease before commencing Neoadjuvant therapy..
in my humble opinion, any lymph node positive disease should undergo resection even with a complete response on imaging since the available imaging methods of assessment of lymph node may miss diseased nodes.. and some people have reservations about giving neoadjuvant therapy in lymph node negative disease...
my review of the studies from exeter and memorial (both retrospective, with small number of patients, median follow up of 26 and 28 months) brings to mind that there can be recurrences if surgery is not performed... no matter how pessimistic and pedantic it may sound, but we will have to await consistent results from high quality randomized trials and their meta-analysis before a practice change occurs..
I agree with Mr Motson that a preferred term would be 'active surveillance' rather than wait and see...
regards, raza
I will proceed to surgery unless very low tumour in young patient who does not go for APE.
Dear Dr.Lin,
Actually I always know, because our practice is to perform full-thickness excision of the scar via transanal endoscopic microsurgery if the endoscopic biopsies are negative for cancer cells.
And the question remains the same.
This is undoubtedly a contentious area but one that is worthy of attention as the comorbidity of low anterior resection is significant and few patients have bowel function as good as they did preoperatively. We have also selectively used the Habr-Gama approach in patients with a complete response in patients with increased surgical risks and have not been 'burnt' so far. I think 'watch and wait' gives the wrong impression of what Angelita (and others including ourselves do) and the close follow up, which is an essential part of this approach is best described as 'active survelillance' with a low threshold to revert to operative intervention if an indication occurs.
@ Dr Julianov: just wondering if TEMS excision of the scar with no mesorectal (lymph node) dissection is enough to declare a pathological complete response..
please correct me if i'm wrong but my impression is that the most of us are referring to complete response on imaging (MRI/EUS/PET-CT etc)..
regards, raza
Dr. Sayyid,
Thanks for asking the question about lymph node response. I have exactly the same question for Dr. Julianov. I am not sure that a complete pathologic tumor response can 100% predict a complete pathologic lymph node response, especially in those patients who need neoadjuvant chemoradiation (at least cT3 or cN1 to start with). Anyone has those data?
Thanks,
Chi
Drs Sayyed and Lin,
Thanks for your comments.
That is why I am asking this question. The uncertainty about N-status is the major and well known problem. Otherwise my question actually will not exist. On the other hand there is evidence from both randomized and nonrandomized studies that in selected patients (regarded as “low-risk” group) chemoradiation followed by local excision is not inferior to standard major surgery. All the problems related with the local treatment are well known and extensively studied and published elsewhere. It is also true that some patients can be spared from unnecessary major surgery and related mortality and morbidity. And I ask this question generally in order to find whether some of you already adopted in everyday practice any kind of selection protocol for ypT0N0M0-patients.
If you look at the history of breast cancer surgery you will see operations like total mastectomy with muscles and internal mammary lymphadenectomy – a procedure we surely cannot agree today. It seems that evolution of the treatment strategy of rectal cancer follows similar way. And some evidence already exist. However most of us, including myself, are very conservative in adopting, or even studying the new concept. I still offer local treatment only to patient who is not fit to surgery or refused it. And I am not satisfied with that.
Fascinating discussion. Thank you Dr Julianov. You are absolutely correct in your observation in how surgical oncology evolves. Breast cancer is a wonderful model for it, but is somewhat confounded by the availability biologically of endocrine therapies that allows more conservative surgery.
Sadly in rectal cancer we do not yet have that chemo therapeutic silver bullet, despite research into biomarkers (my research looks at this) and until we do the risk of local recurrence, and a well recognised aversion of surgeons to change (including myself) we will still have hesitation in recommending local therapy. The breast surgeons are leaps and bounds ahead of us.
I feel it is also important to recognise that as much as we try to make it so, what we do is not a science, but an art. Our patients are not Newtonian predictable machines and what works for some may not work for others. Until we have our patient's and their tumour's genome sequenced (possible and happening in the States) we will have to admit out infancy in knowledge and hold up our hands and say honestly to our patients "We just don't know" but experience, wisdom and common sense, suggests we recommend therapy a or b to you as an individual.
This Philosophical position is beautifully summed up by my favourite saying of Voltaire "Doubt is an unpleasant condition, but certainty is an absurd one"! We must not be precociously arrogant and think we have the answers - there are too many variables and our foundations of medicine and evidence based medicine are not rooted in axiomatic logic, hence making them fallible epistemologically.
Thanks for an exciting discussion thread.
Ironically after all this I feel we're still none the wiser as to what to do!
BW,
JB.
Dear Dr. Julianov,
Thanks for your thoughts and answer.
I actually agree with you on less surgery if delaying the definitive cancer surgery will not compromise the patients' survival or quality of life. It is very reasonable to do so for those with ypT0 cN0 cM0 disease. It really depends on how well we can salvage these patients. I actually think that it is a very good research question for a phase I/II clinical trial.
Thanks,
Chi
When you have a complete clinical response and perform a transanal excision several problems can occur.
First: After a clinical complete response on the staging T the risk of lymph node involvment remains about 6%. The diagnosis of this involvement will be made late. (the chances of cure will be radically different)
Second: If the histologic analysis doesn't confirm the ypT0. You have to perform a complete mesorectal excision, more difficult because of this transanal dissection. Morever the quality of mesorectal dissection will be poor, and this quality has a pronostic impact.
The main problem is that the probability of ypT0N0 after radiochempotherapy is only 15%. This number is too low and the imaging are not effective to diagnostic preoperatively this staging. The risk of mistake in the staging is also too important.
For the anal carcinoma the probability of complete response is almost 75%. With this number it's reasonnable to avoid the surgery.
If we only select pretreatment cT3 cN0 cM0 patient population, when there is a post-chemoRT complete pathologic response at the primary tumor site, the chance that chemoRT (45-50 Gy) has taken care of the microscopic disease in the lymph node is high. I will not feel comfortable to use this approach for patients with pre-chemoRT cN+ disease, because 45-50 Gy is usually considered to be enough for microscopic disease only. Of course, it should be done only in the clinical trial setting.
Anal cancer is different. It is usually squamous cel carcinoma. We give much higher dose (54-59.4 Gy) and with different chemotherapy.
Chi
Thank you Dr Julianov for the explanation..
It is an important question that we will be facing more and more commonly.. and the approach will depend on the patient's and the surgeon's perspective of the risk/benefits of both approaches.. i would like to re-emphasize the point made above that 'watch and wait' or 'wait and see' conveys a rather different impression both to the community of surgeons/oncologists as well as to the patients.. it has to be 'active surveillance' until we get better evidence in this matter...
regards, raza
An interesting phase II english study has assessed the oncologic result of tumoral transanal excision for cT2N0 after radiochemotherapy. However we knows that the radiochemotherapy is not oncologically useful after a mesorectal excision in this situation. Morever the functionnels results are worst after radiochemotherapy. It s not sure that the oncologic and functionnal result are better between a local excision and radiochemotherapy versus a radical surgery alone. It Will be very difficult to respond to this question. The number of patient for a phase III will be too important
Dear Dr. Damont,
I would not say that chemoRT is not useful after total mesorectal excision(TME). A Dutch randomized trial showed that 2 year local recurrent rate decreased from 8% to 2% when combined preoperative chemo RT with TME. (Kapiteijn E, N Eng J Med. 2001 Aug 30; 345(9):638-46.)
Thanks,
Chi
Dear Drs. Dumont and Lin,
Actually there are publiched results from few randomised trials on the subject as this one:
http://www.ncbi.nlm.nih.gov/pubmed/22864880
Dear Dr Julianov
You have righ, your reference is the study that i discuss (and critic) in my previous comment.
Thank you
Merci Dr. Dumont,
En fait, cet article est italien, bien que il a été publié en BJS :)
Bravo pour votre Français. Je suis malheureusement incapable de vous répondre en bulgare.
Merci bien pour votre intéressante question qui fait beaucoup discuter
Cher Dr Dumont, vous êtes très gentil, mais je pense qu'il est juste de continuer en anglais.
Let me to propose an interesting article just published ahead of print in Ann Surg :
http://journals.lww.com/annalsofsurgery/Abstract/publishahead/Survival_Outcome_of_Local_Excision_Versus_Radical.98349.aspx
Dear Drs. Julianov and Dumont;
I assume that you are talking in French which I don't understand. But I do understand that we have data for cT2cNocM0 with this approach. The question is if we can extend this practice to locally advanced disease which would normally require neoadjuvant chemoradiation followed by definitive cancer surgery?
Thanks,
Chi
Hi Chi,
It was just a friendly joke.
Look at these papers:
http://www.ncbi.nlm.nih.gov/pubmed/23849272
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3222828/
http://link.springer.com/article/10.1245%2Fs10434-007-9732-x
http://www.ncbi.nlm.nih.gov/pubmed/15519780
And it will be interesting to see also the long-term results of ACOSOG Z6041 trial:
http://www.ncbi.nlm.nih.gov/pubmed/21755378
The concept clearly is evolving step by step.
Morever, a French multicentric prospective comparative phase III study has begin 4 years ago. This study compare for low rectal cancer T2 or T3 with a good response to radiochemotherapy a local excision versus a TME. The end point is a combination of oncologic and functionnal result. This end point is complicated but It s the only end point able to respond to this question.
Wait a little bit more.
Dr Julianov,
Thank you for posting the references.. i think there is enough equipoise among some of the surgeons about a specific subset of patients that we'll encounter more and more often (unlike the 26 or so patients retrospectively seen over many years in one of the studies), a high quality RCT should be possible and will shed more light on this issue..
as we will encounter this problem more frequently, perhaps there is a need to redefine or subcategorize the pathological complete response into tumor response and tumor+lymph node response..
I congratulate you for generating such an interesting and insightful discussion..
regards, raza
Thanks Raza.
Chi: not all studies are retrospective. Could you agree with this one:
http://www.ncbi.nlm.nih.gov/pubmed/17943364
Yes, Alexander, you are right that it is a phase III trial. I don't disagree that for G1-2 and cT2N0cM0 disease, if neoadjuvant chemoradiation is given, transrectal resection is adequate. But the trial did not include patients with localy advanced disease such as cT3cN0cM0 disease. So, we still don't have data from prospective trials on those patients who will normally receive neoadjuvant chemoradiation because of their locally advanced disease.
Thanks,
Chi
It is true Chi. But just try to extrapolate the number of stage I patients from the CRC incidence datasets and further keep in mind that they are not indicated for neoadjuvant CRT today. That is the reason to start this discussion.
It s true that additional studies are required because this study is a phase II, not a phase III. The only conclusion of this study is that a phase III is interesting to do.
Thank you Dr Julianov for sharing this insightful study..
While the results are very interesting from a randomized controlled trial, we will have to be cautious in interpreting the findings... I am sure everyone agrees that before drawing any conclusions, we must ascertain that the trial was of high quality..
Following were my observations from reviewing the article...
- Method of randomization mentioned is "casual".. i must admit my ignorance for this form of randomization and at least i'm not alone since google also seems unaware of this method of randomization...
- When two surgical interventions are compared in an randomized controlled trial, we have to make sure that both interventions are standardized.. the technique, steps, who performs the operation etc are explicitly mentioned.. i could not find this information..
- If the difference is small or not significant from a comparison of 35 patients in each arm, we cannot conclude that the interventions are equivalent.. i would rather carry the impression that the study is underpowered to detect any difference.
- 23% of laparoscopic resections requiring blood transfusions.. many surgeons who routinely perform laparoscopic procedures for rectal cancers will not be happy with these results.. we can probably say the same about median length of stay of 7 days following lap resections in the era of fast-track surgery... these results need further explanation that i could not find...
- There is no mention of primary outcome measure, but lets assume it is one of the three 'local recurrence, distant mets, probability of survival'.. we don't find a direct comparison, no tables, no p values.. only confidence intervals are mentioned, which are very wide...
- If randomization is adequately performed, one would expect a normal distribution, still medians were used rather than means for almost all contiuous variables.. i failed to find a reason why..
- Wilcoxon test used in table 1 for age and follow-up is for dependent variables, both are certainly not dependent so it is inappropriately used.. reason not mentioned.. similarly table 2 shows wilcoxon for operative time and blood loss.. why??
Based on the above, i have no reservations in saying that this study does not help in making a case for local resections compared to TME..
regards, raza
@Dr Dumont,
My understanding is that if a new intervention is compared against an established form of treatment, then it is phase 3 trial.. phase 2 is not a comparative study and is performed to see the effectiveness and safety of an intervention in a group of patients..
regards, raza
http://www.nlm.nih.gov/services/ctphases.html
It' s true that a phase II assess the safety but you can performed a prospective radomized study in a phase II study. To perform a real comparative phase III study you must make statistical hypothesis to prove the superiority or the equivalence between the arm, and the number of patient in this situation has to be clearly more important. The only conclusion of this study is that the local recurrence and morbidity of the transanal excision after a T2N0 +RtCT is not too important and that a phase III (the statistical hypothesis can be make with the result of this phase II) is interesting to do
Dear Dr Dumont,
Thank you very much for the explanation. I take your point about the phase 2 design..
regards, raza
I love this! While I was sleeping, you guys were having a heated discussion!
I have to confess that I did not read the paper, instead, I only went through the title and the abstract and thought that a prospective randomized study comparing a new method to the standard of care fits the definition of a phase III trial.
I finally got a chance to read the method session of this article and I had to agree that they did a very poor job in terms of study design. It doesn't matter whether it is a phase II or phase III trial, hypothesis and a proper hypothesis driven sample calculation are needed. Of course, the hypothesis is different between phase II and phase III.
The typical hypothesis/objective for phase III will be like this:
"The main objective of the trial was to detect an absolute improvement of (certain number) percent with the control group given a (any endpoint such as 5-year overall survival rate) of (certain number) percent in the group assigned to the experiemental group (with a type I error of 0.05 and a type II error of 0.10). This design required the enrollment of (certain number) patients over a period of certain years."
The hypothesis for phase II will be using historical data for control.
Given that fact that there is no hypothesis driven sample calculation, this study is not standard for either phase III or II.
Besides, randomization could never be done "casually" for either phase III or II.
So, this study at most can be classified as a study on prospectively collected data of 70 consecutively treated patients. I would not say this is either a phase II or phase III study.
Chi
I agree Dr Lin that there are concerns regarding the methodological quality of this study..
in any case, it would require at least a well designed, methodologically sound and adequately powered RCT to answer this important question.. this prospective study of local resections and a comparison to TME may generate enough equipoise to pave the way for such a trial...
regards, raza
Dear Dr lin
It s not the subject and i dont extend the discussion but you have wrong. Search in pubmed "prospective randomized phase II" you have more than 2000 references
Wow! What a strong reviewers! :) Do not focus at single trial, and there are many problems when you try to start surgical RCTs that does not exists in drug RCTs for example. This is a well known and discussed problem in surgical community. If you are interested look at the paper below:
http://journals.lww.com/annalsofsurgery/Abstract/2013/08000/Trends_in_Worldwide_Volume_and_Methodological.1.aspx
May I ask why you agree without any evidence with the current practice to not offer neoadjuvant chemoradiation to patients with stage I and IV rectal cancer? In RCTs of neoadjuvant chemoradiation those patients usually met the exclusion criteria, so the current practice is not evidence-based.
However our topic is different I think.
Dear Dr. Dumont:
You are right that phase II trial can be randomized, which is one of several popular phase II trial design options. Phase II randomized study design usually randomize 2 or more experimental regimens. The investigators will pick the winner and then subject the winning regimen to a definitive phase III trial against standard. Sometimes, you can see that they include a reference standard treatment control arm. but standard control arm acts as a check, not to be compared directly with the experimental arms which is quite different from phase III trials. Randomized phase II design generally requires more patients than a single-arm phase II trial but reduces selection bias and improves patient comparability.
So, the reason that that particular study doesn't qualify for a phase II trial is not because it was a randomized trial, instead it was a study without clear hypothesis and hypothesis driven sample calculation. Besides, randomizing patients "casually" is totally not acceptable for either randomized phase II or III trials.
I know that it is off topic but it is important since we are quoting this to support our arguments.
To answer your question, Alexander, we do have data to support our current practice to not offer neoadjuvant chemoradiation to patients with stage I and IV rectal cancer.
Let's start with stage I,
some early studies on patterns of failure after surgery alone (http://www.ncbi.nlm.nih.gov/pubmed/?term=6192900; http://www.ncbi.nlm.nih.gov/pubmed/?term=6692324; http://www.ncbi.nlm.nih.gov/pubmed/?term=6863077)
have shown that patients treated with surgery alone, the overall risk of local regional recurrence is about 5-15% for stage I, 20-30% for stage II and 20-50% for stage III. These data have been used to justify the use of adjuvant radiation therapy for stage II and III disease but not stage I.
In terms of stage IV, historically data have shown that they are incurable disease and we do palliative treatment. Recently, new data suggested that certain stage IV patients can be cured and we do treat them with neoadjuvant chemoradiation therapy as if they were having locally advanced disease, when the M1 disease is under control.
My point is that we are doing evidence-based practice. At least we are trying to do so or we are extrapolating data from the related existing data.
Again, it may be off topic but important to be cleared.
Chi
Dear Dr Lin,
I agree with the way you have described the available evidence regarding locoregional failure following surgery and hence need for systemic therapy for stage 2 and 3 cancers.
I would just like to add that these figures are from pre-TME era, and with such an established role of neoadjuvant and adjuvant therapy, we will not be able to replicate these studies to see how TME alone has affected the difference in local failure for stage 1, 2 and 3 cancers...
perhaps Dr Julianov is referring to the use of neoadjuvant therapy to stage 1 low rectal cancers with an intention to preserve the sphincters..
regards, raza
Dear Dr. Sayyed,
Thanks for your explanation and addition. I agree that with TME, the local regional failure will be less that that in the pre-TME time. I also agree that we are accumulating more and more evidence on local resection plus neoadjuvant chemoradiation for stage I and perhaps selective stage II patients. However, the change of practice usually takes many years after the evidence is initially presented. I guess that it takes time for physicians to accept new evidences and to discard old practice patterns.
Chi :-)
Of course there should be surgery because of the radiologic limitations highlighted by other and the high relapse rate in those studies where no surgery was performed.
By the way, this thread has been going on for a while, and I think there are almost 80 answers now. What would you think about collecting all this information and write a very original "online poll" opinion paper using ResearchGate?
I think it might be a first...
Thanks @Alexander and to each one: an excellent question and discussion.
At least to me, the following points are of significance:
(1)
It seems not enough concentrating on N-stage only, as lymphatic vessel invasion as well as vascular invasion needs being taking into account (see below).
(2)
We may remind us, that there is a primary bias of 10–25% with regard to loss of sensitivity in the correct determination of the pretherapeutic cN stage only, this bias might be increased afterwards.
(3)
For judging if staging is appropriate the following variables need to be known:
- Sensitivity,
- Specificity,
- Positive predictive value (PPV),
- Negative predictive value (NPV) and
- Overall accuracy
In terms of the performed receiver operating characteristic (ROC) analysis someone may further be informed how this ROC nalysis was performed (which goldstandard was analyzed). If we are hones and demand these variables the available papers providing each variable decreases significantly.
We did this once for evaluating the response in upper GI and it was frustrating as it came up (Table 1 in paper), that some two-thirds of the available publications did “NOT” provide all necessary information [Ann Surg Oncol 2009; 16(4):878-886]. I strongly assume this will be not different in terms of colorectal cancers.
Web link to download PDF:
https://www.researchgate.net/publication/23981159_Response_to_preoperative_therapy_in_upper_gastrointestinal_cancers
(4)
Further the variables above – at least from my perspective - need being known for all different pre-therapeutical stages:
a pre-therapeutic staging of each rectal tumor:
cT2, N0, LVI0, VI
cT2, N0, LVI0 VI1
cT2, N0, LVI1 VI0
cT2, N0, LVI1 VI1
cT2, N1, LVI0, VI
cT2, N1, LVI0 VI1
cT2, N1, LVI1 VI0
cT2, N1, LVI1 VI1
cT3, N0, LVI0, VI
cT3, N0, LVI0 VI1
cT3, N0, LVI1 VI0
cT3, N0, LVI1 VI1
cT3, N1, LVI0, VI
cT3, N1, LVI0 VI1
cT3, N1, LVI1 VI0
cT3, N1, LVI1 VI1
etc
(5)
and they need being known for the post therapeutic staging:
ycT0N0 LVI0 VI0
ycT0N0 LVI0 VI1
ycT0N0 LVI1 VI0
ycT0N0 LVI1 VI1
ycT0N1 LVI0 VI0
ycT0N1 LVI0 VI1
ycT0N1 LVI1 VI0
ycT0N1 LVI1 VI1
etc.
This also helps getting (again) ‘an impression’ about the important point, mentioned before, “how proving a complete response”.
(6)
In terms of this topic, someone may be interested reading a more detailed view of the WHO classification and its points of criticism [Tumor response criteria: are they appropriate?, Fut Oncol 2012;8(8):903-906]:
Page 1:
„Consecutively, the analysis showed that there was an error of 25% in the measurement of the size of identical spheres in 25% of the measurements, and that an error of at least 50% occurred in 6.8% of the measurements. Therefore, using a 25% decrease in tumor size was per definition a criterion for tumor response because it would lead to an unacceptably high rate of response in 25% of cases when in fact the tumor was unchanged. A decrease of 50% was considered to be acceptable for assessing the clinical response, as there would be a low false-positive rate of 6.8%”.
Page 2:
“However, the classification published by Miller et al. in 1981 [9], which was based on a single experiment, was recommended as the definition for response in clinical studies. This was more than 30 years ago. In 2000 – more than 20 years after the first ‘response experiment’ – the US National Cancer Institute, with the European Association for Research and Treatment of Cancer, proposed ‘new response criteria’ for solid tumors; a replacement of 2D measurement with measurement of one dimension was made [11]. Tumor response was defined as a decrease in the largest tumor diameter by 30%, which would translate into a 50% decrease for a spherical lesion [12]. However, no subsequent standardized evaluation of this recommendation was carried out, and 35 years after the primary experiment, no additional studies with a structured logistical way of accurate objective assessment of treatment response have been conducted or proposed”.
Web link to download PDF:
https://www.researchgate.net/publication/230677579_Tumor_response_criteria_are_they_appropriate
Article Response to Preoperative Therapy in Upper Gastrointestinal Cancers
Article Tumor response criteria: Are they appropriate?
I agree in 100% with opinion dr Zoppoli Gabriele. We haven't a prospective trials and how examined CR? EUS, MRI are not adequate in my opinion.
Most patients with a dramatically good response at re-assessment MRI plus photo-endoscopy, (mr TRG 1 & 2) probably best performed at around 10 weeks after CRT should be offered the opportunity of further surveillance .... this must be formal and reliable with extended follow-up with full explanation and support.
- Badges
- Science topic
- Similar topics
- Geoscience
- Asia
- India