Reply

Flynn et al.: does the interval from primary surgery tochemotherapy influence progression-free survival inovarian cancer?

Flynn et al. [1] did a retrospective evaluation of fourplatinum-based chemotherapy trials in ovarian cancer andcould find no evidence from a multivariate analysis that theinterval from surgery to chemotherapy independently influ-enced progression-free survival. They found no contraryreport in the literature, but missed a relevant GynecologicOncology Group (GOG) study [2] of platinum-based com-bination chemotherapy in 349 evaluable cases with smallvolume (residual lesions 1 cm or less) stage III ovariancancer. In multivariate analysis, time from surgery to initi-ation of chemotherapy (which ranged from 1 to 6 weeks)was one of the independent predictors of overall survival(but not of progression-free interval).

One obvious difference is that the majority of the casesexamined by Flynn et al. had larger volume residual disease(2 cm or more). To the extent that early treatment is impor-tant, it may require small volume residual disease to dem-onstrate that effect. Since the question of treatment delaywas not prospectively an objective of the GOG study, andsince the progression-free interval was not affected, theimportance of this finding can be questioned. Nevertheless,there is a strong case to be made from animal modelsdealing with the timing of chemotherapy, that delayingtreatment compromises the outcome [3,4].

Rather than reassuring clinical trialists that delayed treat-ment is acceptable, we should be encouraging them to lookmore closely at the question, and find conditions under whichto prospectively study a potentially important issue.

References

[1] Flynn PM, Paul J, Cruikshank DJ. Does the interval from primarysurgery to chemotherapy influence progression-free survival in ovariancancer? Gynecol Oncol 2002;86:354–7.

[2] Omura GA, Bundy BN, Berek JS, et al. Randomized trial of cyclo-phosphamide plus cisplatin with or without doxorubicin in ovariancarcinoma: a Gynecologic Oncology Group study. J Clin Oncol 1989;7:457–65.

[3] Schabel FM. Rationale for adjuvant chemotherapy. Cancer 1987;39:2875–82.

[4] Griswold DP. Body burden of cancer in relationship to therapeuticoutcome: consideration of preclinical evidence. Cancer Treatment Rep1986;70:81–6.

George A. Omura, M.D.Professor Emeritus, Medicine and Gynecology

University of Alabama at Birmingham3621 Crestside Road

Birmingham, AL 35223, USAE-mail address: geoaomura@aol.com

doi:10.1016/S0090-8258(03)00283-X

Reply

To the Editor:

We are grateful to Dr. Omura for drawing our attentionto this information. As he points out in his letter, the dif-ference in residual disease status between the two studiesmay go some way toward explaining these apparently con-tradictory findings. Re- analysis of a subgroup of our pa-tients (n � 42) who were comparable to those in the GOGstudy has shown a survival hazard ratio of 0.98 compared to1.10 (for every extra week of delay). As the confidencelimits for this hazard ratio in our subgroup (95% C.I. 0.90–1.06) overlap with those in the GOG study (95% C.I. 1.02–1.18), these results do not contradict the GOG findings, butsuggest that if there is an effect of time to surgery onprogression-free survival it is relatively small, as long as thedelay is not too great.

The Gompertzian model of cell kinetics suggests thatthe smaller the volume of a tumour, the greater theproportion that is in growth phase and thereby susceptibleto cytotoxic therapy. Although this results in a greater logkill, unless the tumour is extremely small, there willalmost always be resistant clones which will then multi-ply and eventually cause the death of the patient. Omu-ra’s findings of an improvement in progression-free sur-vival but no influence on overall survival are consistentwith this hypothesis. Whether this results in a net benefitto patients is not clear from either study but quality of lifedata from subsequent studies by our group may allow usto investigate this issue.

Dr. Omura disagrees with our conclusion that the intervalto the start of chemotherapy should not be a major consid-eration for trial organizers. His study confirms that, at leastin terms of overall survival, this is a reasonable position. Weagree that the best course would be to investigate the ques-tion prospectively but given that patients are unlikely towant to be randomized into a study arm that involves de-laying their treatment, such a study may never happen. Theinterval that has elapsed between Dr. Omura’s study andours speaks for itself.

Paul Flynn*Jim Paul

Derek CruickshankSingleton Hospital

Swansea, SA2 8QAWales, UK

E-mail address: paul@obsngobs.com (P. Flynn).

doi:10.1016/S0090-8258(03)00284-1

* Corresponding author. Fax: �44-7092-173892.

498 Letters to the Editor / Gynecologic Oncology 90 (2003) 491–500