In reply to Dr Courdi

We very much thank Dr Courdi for detecting the mistakes and uncertainties in the table in the editorial [1] and in the article by Brooks et al [2]. It is our responsibility to detect mistakes in submitted and accepted papers and definitely not to in an inattentive way replicate them. I apologize for this.

Courdi [3] also brings up two important issues, namely the biologically equivalent doses in 2 Gy fractions concerning acute and late effects to 5×5 Gy, or to the other short term schedules [2], [4] and the α/β ratio for rectal cancer. In his table, Courdi gives the equivalent dose for the late effects. We have still limited knowledge of this dose, but also of the equivalent dose for acute effects in spite that the 5×5 Gy schedule has been used for several decades. When it was originally designed, the old CRE formula was used to estimate the fraction size that, with 5 fractions during 1 week, gives a total dose corresponding to 46 Gy in 2 Gy fractions given 5 days per week. The estimate gave 5.1 Gy, used in the first trials, but later simplified to 5.0 Gy [5].

In an attempt to estimate the α/β for rectal cancer, we have analysed the 15 randomized trials comparing preoperative radiotherapy and surgery with surgery alone with complete data, identified by the Colorectal Cancer Collaborative Group [6] together with the Dutch TME trial (Glimelius and Isaksson, unpublished observations). Based upon the relative reduction (RR) in local failure rates in the radiotherapy group compared to the surgery alone group in the trials, we aimed at estimating the parameters in the LQ formula by minimizing the sum of weighted square differences between the RR and the model expression. The best fit in the LQ model was seen with an α/β = 3.9 (±95CI 0.6) Gy (correlation coefficient 0.62, p < 0.05). If a time corrected LQ model was used, the best fit was seen when α/β = 2.7±0.3 Gy and γ/α = 1.6±0.3 Gy/day (correlation coefficient 0.78, p < 0.01). Using these parameter values, the equivalent doses in 2 Gy fractions of 5 x 5 Gy correspond to about 36 Gy and 76 Gy, respectively. If the γ/α is set to 0.6 Gy/day, revealing only slightly less fit to the data (r = 0.76), the equivalent dose is about 50 Gy. These estimations, based upon the available clinical data from the randomized trials, firstly tell that the assumption by Courdi, namely that the α/β ratio for rectal cancer is rather low appears correct. Secondly, they show that we have very little knowledge about the equivalent antitumour dose. The ongoing Stockholm III trial comparing 5×5 Gy with immediate or delayed surgery and 25×2 Gy with delayed surgery (Cedermark B, personal communication) may give a better answer, although the Polish trial [7] indicates that it is likely more than 50 Gy.