Acute toxicity sub-studies

Background

Acute or short-term effects of RT occur during and shortly after RT. Acute effects are commonly reddening of the skin, usually mild, but occasionally leading to severe redness, or blistering of the skin. They can also include discomfort, swelling and fatigue. The severity of acute side effects is related to total dose given, rather than dose per fraction. In FAST-Forward, the severity of acute side effects was not expected to be worse with the 1-week schedules, but actually to be milder, and shorter-lived due to the lower total dose. This was supported by previous studies such as START and FAST, where in a subset of patients, acute toxicity was evaluated and found to be milder in patients having hypofractionated schedules, compared to the standard 50 Gy in 25 fractions. However, despite this, many clinicians and patients assume that the higher dose per fraction would lead to much more severe reactions, quoting that it would cause severe reactions, or ‘burns’. It was therefore decided to carry out a detailed assessment of acute side effects in the first cohort of patients recruited at a subset of centres participating in the FAST-Forward Main Trial.

Methods

Patients and centres

The acute toxicity sub-studies were conducted in a subset of centres (Royal Marsden Hospital, Royal Cornwall Hospital, University Hospitals of North Midlands, Derriford Hospital Plymouth, Norfolk and Norwich University Hospitals and Addenbrookes Hospital) that had the infrastructure necessary to carry out weekly toxicity assessments. The same centres participated in both sub-studies, with the addition of Torbay Hospital in the second sub-study. Patients receiving a RT tumour bed boost were eligible for the first acute toxicity sub-study but were excluded from the second acute toxicity sub-study since the objective of this sub-study was to quantify the toxicity of 5-fraction schedules relative to control, effects that are independent and additive to those of the boost. At centres participating in the acute toxicity sub-studies, consecutive eligible patients were invited to participate at the time of consent to the Main Trial.

Assessments

Acute reactions of the skin of the treated breast were graded using RTOG criteria for the first sub-study (Protocol version 1.0). Following discussions with both the FAST-Forward IDMC and TSC, it was agreed that this may not be the most appropriate scoring system in this context due to the inclusion of oedema in the RTOG system, since the primary concern was to score moist desquamation. Therefore, it was agreed that a second sub-study would be undertaken using standard common toxicity criteria for adverse effects (CTCAE) criteria (v4.03) (Protocol versions 2.1 and 2.2). Toxicity assessments were made by a healthcare professional at each participating centre using predesigned forms. The assessments were scheduled to be carried out weekly during treatment and for 4 weeks following the end of RT. In the second sub-study, if moist desquamation beyond skin folds or creases (CTCAE grade 3) were seen during this time, weekly assessments were to continue until the reaction resolved to CTCAE grade 1 or less. If any assessment was missed, the centre contacted the patient by telephone to assess and grade acute skin reactions by asking the patient to describe the skin appearance and supplementing this by direct questions. In the first acute toxicity sub-study patient self-assessments of acute AE (breast soreness, reddening, swelling and blistering) were recorded using patient diary cards completed weekly during RT and for at least 4 weeks post-RT. The scores were recorded as ‘none’, ‘a little’, ‘quite a bit’ or ‘very much’. If symptoms persisted, patients were asked to continue scoring their AE on a weekly basis until all scores were graded as ‘none’ or ‘a little’. The submitted patient diaries showed low data completion to the extent that the results were not considered of value to be included in the submitted manuscript.

Acute skin reactions scoring scales

Modified RTOG scale (acute toxicity study I)

• Grade 0: No visible change.
• Grade 1: Faint/dull erythema.
• Grade 2: Tender/bright erythema ± dry desquamation.
• Grade 3: Patchy moist desquamation, moderate oedema.
• Grade 4: Confluent moist desquamation, pitting oedema.

CTCAE version 4.03 (acute toxicity study II)

• Grade 1: Faint erythema or dry desquamation.
• Grade 2: Moderate to brisk erythema; patchy moist desquamation, mostly confined to skin fold and creases; moderate oedema.
• Grade 3: Moist desquamation in areas other than skin folds and creases; bleeding induced by minor trauma or abrasion.
• Grade 4: Life-threatening consequences; skin necrosis or ulceration of full thickness dermis; spontaneous bleeding from involved site; skin graft indicated.
• Grade 5: Death.

Endpoints

The primary endpoint for each of the sub-studies was the proportion of patients within each treatment group with grade ≥ 3 toxicity (RTOG and CTCAE, respectively) at any time from the start of RT to 4 weeks after completion of RT. The secondary endpoints were clinical assessment of (1) any acute skin toxicity, defined as the worst grade reported from the start of RT to 4 weeks post-RT, (2) acute skin toxicity during RT, defined as the worst grade reported from the start to the end of RT, (3) acute skin toxicity post-RT, defined as the worst grade reported at completion of RT until at least 4 weeks post-RT and (4) adherence to the acute toxicity assessments.

Statistical considerations

Each sub-study aimed to recruit approximately 150 evaluable patients (50 per group) in order to exclude a within-group rate of grade ≥ 3 acute skin reactions of over 11% compared to target rate of under 3% (89% power and one-sided 7.9% significance level). The sample size was informed by data from the FAST pilot trial.26

Principal analyses were based on the evaluable population, defined as all patients randomised into the study receiving at least one fraction of RT (regardless of whether they were later found to be ineligible or a protocol violator) and with complete data, or, at most, one missing toxicity assessment. For the first sub-study which included some patients receiving a boost, only non-boost assessments were used to define inclusion in the evaluable population. The required number of assessments for a 40 Gy/15 fractions patient was seven (three during and four post-RT) and five for 27 Gy/5 fractions and 26 Gy/5 fractions patients (one during and four post-RT). Given the focus of the sub-studies on safety patients who switched to receive a different trial treatment after randomisation were analysed according to treatment received rather than randomised treatment.

For the primary endpoint, the proportion of patients within each treatment group with grade ≥ 3 RTOG toxicity (first sub-study) or grade ≥ 3 CTCAE toxicity (second sub-study) was estimated with associated upper one-sided 95% CI. Secondary endpoints were estimated as frequencies and percentages, with the prevalence of grades ≥ 1, ≥ 2 and ≥ 3 toxicity at each time point presented graphically. Adherence to toxicity assessments was estimated as frequencies and percentages. Toxicity data were included as reported, and there were no restrictions based on the date of assessment. The acute toxicity sub-study was designed to be non-comparative so no statistical comparisons were made across the treatment groups.

Results

Recruitment to the first acute toxicity sub-study opened in November 2011 and closed in May 2012. Of 194 potentially eligible patients recruited in sub-study centres, 190 patients consented (63, 63 and 64 in the 40 Gy/15 fractions, 27 Gy/5 fractions and 26 Gy/5 fractions groups, respectively). Recruitment was performed at seven treating centres across the UK with the infrastructure to carry out weekly toxicity assessments.

Recruitment to the second acute toxicity sub-study opened in April 2013 and closed in February 2014. Of 269 potentially eligible patients recruited in the 8 sub-study centres, 162 patients consented (55, 44 and 63 in the 40 Gy/15 fractions, 27 Gy/5 fractions and 26 Gy/5 fractions groups, respectively). Two patients subsequently withdrew consent for all data to be used. The number of patients potentially eligible, consented, receiving allocated RT and evaluable for the primary endpoint for each sub-study is given in Figure 3.

FIGURE 3 Flow diagram of participation in the first and second acute toxicity sub-studies. Reproduced from Brunt et al. (CC BY 4.0).3 aAll patients recruited at sub-study centres during sub-study period; bAll patients recruiting at sub-study centres during sub-study period for whom a boost was not planned. (Figure not reproduced here; see the original article.)

Patients recruited to the sub-studies were broadly comparable with those of the whole trial population.3 Radiotherapy treatment characteristics (for each sub-study separately) are presented in Appendix 2, Table 36, suggesting no significant imbalances between groups. Compliance to toxicity assessments by treatment group is shown in Appendix 2, Table 37. For the first acute toxicity sub-study, worst RTOG grade experienced by treatment received is reported in Table 2. The proportions of evaluable patients with grade 3 RTOG toxicity during the acute phase were as follows: 40 Gy/15 fractions 6/44 (13.6%); 27 Gy/5 fractions 5/51 (9.8%); 26 Gy/5 fractions 3/52 (5.8%). Twenty-nine patients in the first sub-study received a boost to their treatment, of which 22 were evaluable. No evidence of a higher rate of grade 3 toxicity was observed in this subset of patients [40 Gy/15 fractions 1/9 (11.1%); 27 Gy/5 fractions 0/10; 26 Gy/5 fractions 0/3).

For the second acute toxicity sub-study, worst CTCAE grade experienced by treatment received is reported in Table 3. The proportions of evaluable patients with a grade 3 CTCAE toxicity during the acute phase were as follows: 40 Gy/15 fractions 0/43; 27 Gy/5 fractions 1/41 (2.4%); 26 Gy/5 fractions 0/53. Grade 2 toxicity was largely due to moderate to brisk erythema, with only three patients with moderate oedema (two 40 Gy/15 fractions and one 26 Gy/5 fractions patient), see Appendix 2, Table 38. CTCAE toxicity score reported at each time point is presented graphically during RT and post-RT (see Figure 4). A single patient in the second cohort randomised to the 27 Gy/5 fractions group after mastectomy, adjuvant cytotoxic chemotherapy and chest wall RT with bolus was assessed as a CTCAE grade 3 toxicity recorded after a missed assessment by telephone appointment 36 days after start of treatment, having been assessed as grade 2 7 days before and assessed with grade 1 erythema 7 days later.

FIGURE 4 Acute toxicity sub-study 2 – prevalence of grade 1+, grade 2+ and grade 3+ CTCAE toxicity. Reproduced from Brunt et al. (CC BY 4.0).3 Source: Grade 3 toxicity reported at 4 weeks post-RT in 27 Gy in 5 fractions patient resolved to grade 1 1 week later. (Figure not reproduced here; see the original article.)

Discussion

The acute skin toxicity sub-studies were not designed to involve statistical hypothesis testing across treatment groups, but to confirm low incidence rates of clinically significant acute skin toxicity associated with each schedule. On this basis, the results raised no concerns that the 5-day schedules led to more severe or longer-lasting acute skin reactions compared with 40 Gy in 15 fractions. The prevalence rates suggested that erythema after the 1-week schedule was less intense and settled about 2 weeks earlier than after the 3-week schedule. An imbalance in numbers in the 27 Gy group was attributed to the play of chance, since consent to participate in the sub-study was obtained prior to allocation of randomised treatment. Overall compliance with the toxicity assessments was high, making it unlikely that patients developing severe skin reactions were under-reported. No suggestion of a dose–response was noted between the two dose levels of the 5-fraction test regimens. The results of the first toxicity sub-study were consistent with those from the second sub-study, despite the unhelpful inclusion of oedema with erythema and desquamation in the RTOG system used for scoring early skin reactions.

The mildness of the acute skin toxicity associated with the 5-fraction regimens was expected. A series of classic studies investigating the dependence of acute skin reactions on total dose, fraction size, interfraction interval and overall treatment time was undertaken in the 1980s and 1990s by Turesson and colleagues, using reflectance spectrophotometry to quantify erythema and clinical grading to score moist desquamation.65–67 These confirmed the absence of a treatment time effect for the first 4 weeks of RT delivered using fraction sizes of 2.0 Gy and 4.0 Gy delivered using 12 MeV electrons to the IMC daily or twice per week.65 Using the same experimental system, whereby different fractionation regimens were delivered to the right and left IMCs of the same patient, weak dependence of erythema and desquamation on fraction size was demonstrated, with estimates of α/β ratios between 7.5 Gy and 11.2 Gy.66 In the context of the FAST-Forward Trial, the reductions in total dose, from 40 Gy to 27 Gy and 26 Gy appear to compensate for increased dose per fraction where acute skin reactions are concerned. In conclusion, the acute skin toxicity sub-studies conducted in patients entered into the FAST-Forward Trial raised no concerns and showed that short-term effects are if anything milder, and shorter lasting.