Sir,
Shabaruddin et al (2010) recently reported on the results of a survey conducted by expert UK oncologists regarding their use by irinotecan for colorectal cancer, with the intent of using these data to explore the cost-effectiveness of routine UGT1A1 testing. Reading this article, we were struck with the wide variation in the experts’ answers, and wondered, as did the authors, as to how accurate the experts might be. To further explore this observation, we put the same questions to seven Melbourne (Australia) oncologists who work at four institutions. We then compared these responses with prospective data that they had collected on 604 patients with metastatic colorectal cancer, who had presented with stage IV disease, or developed recurrence, between 2005 and 2009.
We were initially struck by the similarity of the estimates made by the UK and Melbourne experts (Table 1) regarding treatment used in the first- to third-line setting. Examining the Melbourne data, it was evident that the Melbourne experts were overestimating the use of irinotecan, with even the lowest estimate of any expert being higher than real practice. In the third-line setting, regarding the use of mitomycin C, the discordance between estimates and real data was most striking (40% estimate versus 17.4% reality). This is consistent with clinicians reporting that they found it more difficult to estimate treatment use with increasing lines of therapy.
The Melbourne experts’ estimates for irinotecan duration in the second-line setting were quite accurate; however, in the first-line setting, in which all but one expert said either 5 or 6 months, the real data showed 3.2 months. The estimates of the incidence of febrile neutropenia ranged from 2 to 8%. Remarkably, the mean result was 5.4%, which is the exact figure found for the real data.
One possible explanation for the similarity between the UK and Melbourne estimates, in instances in which there are comparable data, is that both groups may be basing their answers on a common set of reference data, such as clinical trial results or other published data. Experts may then use this knowledge to derive their estimates, rather than have a true awareness of what is happening in their practice. To explore this possibility further, we compared the Melbourne experts’ estimates with available trial data. The duration of irinotecan use in first-line trials was 5.5–6.5 months (Douillard et al, 2000; Saltz et al, 2000; Tournigand et al, 2004) and that in second-line studies was 1.5–4.1 months (Cunningham et al, 1998; Tournigand et al, 2004; Sobrero et al, 2009; Gibbs et al, 2010), which was quite similar to the experts’ estimates (see Table 1). In two studies enrolling patients in a first-line setting, the percentage who then received second-line treatment was reported as 54 and 62% (Tournigand et al, 2004; Seymour et al, 2007), which is again similar to the experts’ estimates. The reported incidence of febrile neutropenia ranged from 2.1% (Douillard et al, 2000) to 7.1% (Sobrero et al, 2009), again in line with the experts’ estimates.
If the above explanation were true, we would also anticipate that experts’ estimates would be accurate when real data were similar to trial data, and inaccurate when, for whatever reason, there were significant differences between clinical studies and real life. This is most obvious for the duration of use of first-line irinotecan, in which further analysis of our hospital treatment data revealed that the dominant use of irinotecan in this setting was when patients had recently failed oxaliplatin-based adjuvant treatment, a poor prognosis group. This would explain the duration of first-line irinotecan treatment being similar to irinotecan use in the second-line setting, wherein patients typically receive treatment following failure of first-line oxaliplatin-based therapy. Although in retrospect this is an obvious explanation for the abbreviated first-line irinotecan-based treatment, it appears that only one of our experts may have factored this into their answers.
Finally, we also suggest that when there were no available trial data as a reference point, estimates were not only likely to be inaccurate but there was also a much wider variation in the answers provided by the experts. An example of such a query would be the use of third-line mitomycin C, or the reasons for discontinuing irinotecan. When these are examined, the data (Table 1) appear consistent with this observation. Estimates of third-line mitomycin C use ranged from 0 to 80%, and therapy completion as the reason for cessation of therapy ranged from 10 to 50%; in both instances the means were quite different from the real data.
It is worth noting that when data are combined, as would be required for the analysis that Shabaruddin et al were undertaking, the inaccuracy of the experts’ opinions across the multiple data sets would be compounded. For example, regarding first-line irinotecan use, the amount of drug used per 100 patients in the Melbourne experts’ opinion would be 66 months (5 months duration in 13.2 patients per 100), a total that is 41.1 months or 3.49-fold greater than the real data (18.9 months=3.2 months for 5.9% of patients). Other estimates would be closer, including second-line irinotecan use, which parallels clinical trial data; others would be out by much more, such as the third-line use of mitomycin C.
Historically, for a speciality so driven by clinical trials that produce large amounts of data, we would agree with Shabaruddin et al that the scant amount of data regarding treatment and outcomes in routine care is surprising. Although collecting data is achievable, as we have demonstrated (see also http://www.biogrid.org.au), we acknowledge that it does require significant effort and resources. However, if understanding what is happening in routine care is important, and we believe it is for multiple reasons, then we would suggest that the necessary investments to support this must be made. On the basis of our small study, using the estimates of experts appears to be of limited value, and could be potentially quite misleading. These estimates may also become increasingly unreliable as treatment paradigms become more complex, use of biological therapy becomes more widespread and multiple lines of therapy become more common.
Change history
29 March 2012
This paper was modified 12 months after initial publication to switch to Creative Commons licence terms, as noted at publication
References
Cunningham D, Pyrhonen S, James RD, Punt CJA, Hickish TF, Heikkila R, Johannesen TB, Starkhammar H, Topham CA, Awad L, Jacques C, Herait P (1998) Randomised trial of irinotecan plus supportive care versus supportive care alone after fluorouracil failure for patients with metastatic colorectal cancer. Lancet 352: 1413–1418
Douillard JY, Cunningham D, Roth AD, Navarro M, James RD, Karasek P, Jandik P, Iveson T, Carmichael J, Alakl M, Gruia G, Awad L, Rougier P (2000) Irinotecan combined with fluorouracil compared with fluorouracil alone as firstline treatment for metastatic colorectal cancer: a multicentre randomised trial. Lancet 355: 1041–1047
Gibbs P, Clingan PR, Ganju V, Strickland AH, Wong SS, Tebutt NC, Underhill CR, Fox RM, Clavant SP, Leung J, Pho M, Brown TJ (2010) Hyaluronan-Irinotecan improves progression-free survival in 5-fluorouracil refractory patients with metastatic colorectal cancer: a randomized phase II trial. Cancer Chemother Pharmacol March 24 (e-pub ahead of print)
Saltz LB, Cox JV, Blanke C, Rosen LS, Fehrenbacher L, Moore MJ, Maroun JA, Ackland SP, Locker PK, Pirotta N, Elfring GL, Miller LL (2000) Irinotecan plus fluorouracil and leucovorin for metastatic colorectal cancer. Irinotecan Study Group. N Engl J Med 343 (13): 905–914
Seymour MT, Maughan TS, Ledermann JA, Topham C, James R, Gwyther SJ, Smith DB, Shepherd S, Maraveyas A, Ferry DR, Meade AM, Thompson L, Griffiths GO, Parmar MK, Stephens RJ (2007) Different strategies of sequential and combination chemotherapy for patients with poor-prognosis advanced colorectal cancer (MRC FOCUS): a randomised controlled trial. Lancet 370: 143–152
Shabaruddin FH, Elliott RA, Valle JW, Newman WG, Payne K (2010) Understanding chemotherapy treatment pathways of advanced colorectal cancer patients to inform an economic evaluation in the United Kingdom. Br J Cancer 103: 315–323
Sobrero A, Ackland S, Clarke S, Perez-Carrión R, Chiara S, Gapski J, Mainwaring P, Langer B, Young S (2009) Phase IV study of Bevacizumab in combination with infusional Fluorouracil, Leucovorin and Irinotecan (FOLFIRI) in first-line metastatic colorectal cancer. Oncology 77: 113–119
Tournigand C, André T, Achille E, Lledo G, Flesh M, Mery-Mignard D, Quinaux E, Couteau C, Buyse M, Ganem G, Landi B, Colin P, Louvet C, de Gramont A (2004) FOLFIRI followed by FOLFOX6 or the reverse sequence in advanced colorectal cancer: a randomized GERCOR study. J Clin Oncol 22: 229–237
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
From twelve months after its original publication, this work is licensed under the Creative Commons Attribution-NonCommercial-Share Alike 3.0 Unported License. To view a copy of this license, visit http://creativecommons.org/licenses/by-nc-sa/3.0/
About this article
Cite this article
Heong, V., Ananda, S., Tie, J. et al. Guesstimates are not good enough for determining what is happening in routine care. Br J Cancer 103, 1885–1886 (2010). https://doi.org/10.1038/sj.bjc.6605998
Published:
Issue Date:
DOI: https://doi.org/10.1038/sj.bjc.6605998
This article is cited by
-
Reply: Guesstimates are not good enough for determining what is happening in routine care
British Journal of Cancer (2010)