Analysis of late toxicity associated with external beam radiation therapy for prostate cancer with uniform setting of classical 4-field 70 Gy in 35 fractions: a survey study by the Osaka Urological Tumor Radiotherapy Study Group.

We aimed to analyse late toxicity associated with external beam radiation therapy (EBRT) for prostate cancer using uniform dose-fractionation and beam arrangement, with the focus on the effect of 3D (CT) simulation and portal field size. We collected data concerning patients with localized prostate adenocarcinoma who had been treated with EBRT at five institutions in Osaka, Japan, between 1998 and 2006. All had been treated with 70 Gy in 35 fractions, using the classical 4-field technique with gantry angles of 0°, 90°, 180° and 270°. Late toxicity was evaluated strictly in terms of the Common Terminology Criteria for Adverse Events Version 4.0. In total, 362 patients were analysed, with a median follow-up of 4.5 years (range 1.0-11.6). The 5-year overall and cause-specific survival rates were 93% and 96%, respectively. The mean ± SD portal field size in the right-left, superior-inferior, and anterior-posterior directions was, respectively, 10.8 ± 1.1, 10.2 ± 1.0 and 8.8 ± 0.9 cm for 2D simulation, and 8.4 ± 1.2, 8.2 ± 1.0 and 7.7 ± 1.0 cm for 3D simulation (P < 0.001). No Grade 4 or 5 late toxicity was observed. The actuarial 5-year Grade 2-3 genitourinary and gastrointestinal (GI) late toxicity rates were 6% and 14%, respectively, while the corresponding late rectal bleeding rate was 23% for 2D simulation and 7% for 3D simulation (P < 0.001). With a uniform setting of classical 4-field 70 Gy/35 fractions, the use of CT simulation and the resultant reduction in portal field size were significantly associated with reduced late GI toxicity, especially with less rectal bleeding.

INTRODUCTION

Since radiotherapy for prostate cancer is a standard treatment option for localized prostate cancer, its toxicity should be clearly addressed. In a previous survey study conducted from 1995 to 2006 in Osaka, Japan, which was intended to clarify time trends in radiotherapy and its biochemical relapse-free survival (bRFS) outcomes, we made the interesting discovery that 87% of patients had been treated with a highly uniform mode of radiotherapy, that is, with classical 4-field 70 Gy in 35 fractions [1]. While that study was being conducted, CT simulation was introduced and developed, and almost all institutions had replaced 2D simulation with 3D simulation by 2006. This resulted in a reduction in the size of the portal field. We realized that we could obtain very pure data for an investigation of the relationship between portal field size and late toxicity rate, especially for rectal bleeding, in view of the uniform setting of dose-fractionation and beam arrangement. The findings of this investigation are the main subject of this article. Dearnaley et al. had already conducted a prospective randomized trial comparing 1.0 and 1.5 cm margins, and concluded that a larger margin was associated with significantly higher incidence of toxicities [2]. However, their study included only 126 patients, who had been assigned to 2 × 2 arms (64 Gy and 74 Gy groups, and 1.0 and 1.5 cm margin groups). Moreover, their treatment planning included two phases comprising a 3-field phase and a 6-field phase. We aimed to repeat the investigation with a larger Japanese patient cohort, treated with more uniform dose-fractionation and beam arrangement, although in a retrospective manner.

MATERIALS AND METHODS

A brief summary of the previous survey study

In our previous study [1], data were collected for 652 consecutive patients with clinically localized prostate cancer (T1-4N0M0), who had been treated with definitive external beam radiotherapy (EBRT) of 60 Gy or more at one of the 11 participating institutions, mainly in Osaka, Japan, from 1995 through 2006. Of the 652 patients, 436 met the enrolment criteria and were analysed. The main findings were: (i) the number of radiotherapy patients showed a 10-fold increase over 10 years; (ii) the dominant dose-fractionation was 70 Gy/35 fractions (87%); (iii) hormone therapy had been administered to 95% of the patients; (iv) the 3- and 5-year bRFS rates were 85% and 70%, respectively; (v) toxicity data was not available.

An interesting finding was that as many as 87% of the patients had received radiotherapy in a highly uniform manner, that is, with the classical 4-field technique using a dose-fractionation schedule of 70 Gy/35 fractions. We therefore planned the second survey by focusing on detailed late toxicity data and irradiation field data obtained with a uniform setting of 4-field 70 Gy/35 fractions.

Data collection

Five institutions participated in the present study. Data collected for the 362 patients who are the subject of this study are described in the Results section.

All the data were collected by physicians (radiation oncologists or urologists), who are also the authors of this paper, and no non-physician surveyors took any part in this study. Detailed information was collected about portal field size and other parameters of radiotherapy. Late toxicity grading was performed by retrospectively reviewing medical charts, strictly according to the Common Terminology Criteria for Adverse Events (CTCAE) Version 4.0. The concrete description of each relevant CTCAE Term was gathered as Tables 1 and 2, which had been distributed to the surveyors as references. All data were sent to Osaka University, analysed by the first author, and finally reviewed and approved by all the authors.

Table 1.

Late genitourinary toxicity scale extracted from the Common Terminology Criteria for Adverse Events (CTCAE) Version 4.0

Renal and urinary disorders
	Grade
Adverse Event	1	2	3	4	5
Hematuria	Asymptomatic; clinical or	Symptomatic; urinary catheter	Gross hematuria; transfusion, IV	Life-threatening consequences;	Death
	diagnostic observations only;	or bladder irrigation indicated;	medications or hospitalization	urgent radiologic or operative
	intervention not indicated	limiting instrumental ADL	indicated; elective endoscopic,	intervention indicated
			radiologic or operative
			intervention indicated; limiting
			self care ADL
Definition: A disorder characterized by laboratory test results that indicate blood in the urine.
Urinary frequency	Present	Limiting instrumental ADL;	–	–	–
		medical management indicated
Definition: A disorder characterized by urination at short intervals.
Urinary incontinence	Occasional (e.g. with coughing,	Spontaneous; pads indicated;	Intervention indicated (e.g.	–	–
	sneezing, etc.), pads not	limiting instrumental ADL	clamp, collagen injections);
	indicated		operative intervention indicated;
			limiting self care ADL
Definition: A disorder characterized by inability to control the flow of urine from the bladder.
Urinary retention	Urinary, suprapubic or	Placement of urinary,	Elective operative or radiologic	Life-threatening consequences;	Death
	intermittent catheter placement	suprapubic or intermittent	intervention indicated;	organ failure; urgent operative
	not indicated; able to void with	catheter placement indicated;	substantial loss of affected	intervention indicated
	some residual	medication indicated	kidney function or mass
Definition: A disorder characterized by accumulation of urine within the bladder because of the inability to urinate.
Urinary tract obstruction	Asymptomatic; clinical or	Symptomatic but no	Symptomatic and altered organ	Life-threatening consequences;	Death
	diagnostic observations only	hydronephrosis, sepsis or renal	function (e.g. hydronephrosis,	urgent intervention indicated
		dysfunction; urethral dilation,	or renal dysfunction); elective
		urinary or suprapubic catheter	radiologic, endoscopic or
		indicated	operative intervention indicated
Definition: A disorder characterized by blockage of the normal flow of contents of the urinary tract.
Urinary tract pain	Mild pain	Moderate pain; limiting	Severe pain; limiting self care	–	–
		instrumental ADL	ADL
Definition: A disorder characterized by a sensation of marked discomfort in the urinary tract.
Urinary urgency	Present	Limiting instrumental ADL;	–	–	–
		medical management indicated
Definition: A disorder characterized by a sudden compelling urge to urinate.
Urine discoloration	Present	–	–	–	–
Definition: A disorder characterized by a change in the color of the urine.
Renal and urinary disorders -	Asymptomatic or mild	Moderate, local or noninvasive	Severe or medically significant	Life-threatening consequences;	Death
Other, specify	symptoms; clinical or diagnostic	intervention indicated; limiting	but not immediately life-	urgent intervention indicated
	observations only; intervention	instrumental ADL	threatening; hospitalization or
	not indicated		prolongation of existing
			hospitalization indicated;
			disabling; limiting self care ADL

Table 2.

Late gastrointestinal toxicity scale extracted from the Common Terminology Criteria for Adverse Events (CTCAE) Version 4.0

Gastrointestinal disorders
	Grade
Adverse Event	1	2	3	4	5
Abdominal pain	Mild pain	Moderate pain; limiting	Severe pain; limiting self care	–	–
		instrumental ADL	ADL
Definition: A disorder characterized by a sensation of marked discomfort in the abdominal region.
Anal fistula	Asymptomatic; clinical or	Symptomatic; altered GI	Severely altered GI function;	Life-threatening consequences;	Death
	diagnostic observations only;	function	tube feeding, TPN or	urgent intervention indicated
	intervention not indicated		hospitalization indicated;
			elective operative intervention
			indicated
Definition: A disorder characterized by an abnormal communication between the opening in the anal canal to the perianal skin.
Anal hemorrhage	Mild; intervention not indicated	Moderate symptoms; medical	Transfusion, radiologic,	Life-threatening consequences;	Death
		intervention or minor	endoscopic, or elective	urgent intervention indicated
		cauterization indicated	operative intervention indicated
Definition: A disorder characterized by bleeding from the anal region.
Anal mucositis	Asymptomatic or mild	Symptomatic; medical	Severe symptoms; limiting self	Life-threatening consequences;	Death
	symptoms; intervention not	intervention indicated; limiting	care ADL	urgent intervention indicated
	indicated	instrumental ADL
Definition: A disorder characterized by inflammation of the mucous membrane of the anus.
Anal necrosis	–	–	TPN or hospitalization indicated;	Life-threatening consequences;	Death
			radiologic, endoscopic, or	urgent operative intervention
			operative intervention indicated	indicated
Definition: A disorder characterized by necrotic process occurring in the anal region.
Anal pain	Mild pain	Moderate pain; limiting	Severe pain; limiting self care	–	–
		instrumental ADL	ADL
Definition; A disorder characterized by a sensation of marked discomfort in the anal region.
Anal stenosis	Asymptomatic; clinical or	Symptomatic; altered GI	Symptomatic and severely	Life-threatening consequences;	Death
	diagnostic observations only;	function	altered GI function; non-	urgent operative intervention
	intervention not indicated		emergent operative intervention	indicated
			indicated; TPN or hospitalization
			indicated
Definition: A disorder characterized by a narrowing of the lumen of the anal canal.
Anal ulcer	Asymptomatic; clinical or	Symptomatic; altered GI	Severely altered GI function;	Life-threatening consequences;	Death
	diagnostic observations only;	function	TPN indicated; elective	urgent operative intervention
	intervention not indicated		operative or endoscopic	indicated
			intervention indicated; disabling
Definition: A disorder characterized by a circumscribed, inflammatory and necrotic erosive lesion on the mucosal surface of the anal canal.
Constipation	Occasional or intermittent	Persistent symptoms with	Obstipation with manual	Life-threatening consequences;	Death
	symptoms; occasional use of	regular use of laxatives or	evacuation indicated; limiting	urgent intervention indicated
	stool softeners, laxatives,	enemas; limiting instrumental	self care ADL
	dietary modification, or enema	ADL
Definition: A disorder characterized by irregular and infrequent or difficult evacuation of the bowels.
Diarrhea	Increase of <4 stools per day	Increase of 4-6 stools per day	Increase of ≥7 stools per day	Life-threatening consequences;	Death
	over baseline; mild increase in	over baseline; moderate	over baseline; incontinence;	urgent intervention indicated
	ostomy output compared to	increase in ostomy output	hospitalization indicated; severe
	baseline	compared to baseline	increase in ostomy output
			compared to baseline; limiting
			self care ADL
Definition: A disorder characterized by frequent and watery bowel movements.
Fecal incontinence	Occasional use of pads required	Daily use of pads required	Severe symptoms; elective	–	–
			operative intervention indicated
Definition: A disorder characterized by inability to control the escape of stool from the rectum.
Hemorrhoidal hemorrhage	Mild; intervention not indicated	Moderate symptoms; medical	Transfusion, radiologic,	Life-threatening consequences;	Death
		intervention or minor	endoscopic, or elective	urgent intervention indicated
		cauterization indicated	operative intervention indicated
Definition: A disorder characterized by bleeding from the hemorrhoids.
Hemorrhoids	Asymptomatic; clinical or	Symptomatic; banding or	Severe symptoms; radiologic,	–	–
	diagnostic observations only;	medical intervention indicated	endoscopic or elective operative
	intervention not indicated		intervention indicated
Definition: A disorder characterized by the presence of dilated veins in the rectum and surrounding area.
Ileus	–	Symptomatic; altered GI	Severely altered GI function;	Life-threatening consequences;	Death
		function; bowel rest indicated	TPN indicated	urgent intervention indicated
Definition: A disorder characterized by failure of the ileum to transport intestinal contents.
Proctitis	Rectal discomfort, intervention	Symptoms (e.g. rectal	Severe symptoms; fecal	Life-threatening consequences;	Death
	not indicated	discomfort, passing blood or	urgency or stool incontinence;	urgent intervention indicated
		mucus); medical intervention	limiting self care ADL
		indicated; limiting instrumental
		ADL
Definition: A disorder characterized by inflammation of the rectum.
Rectal fistula	Asymptomatic; clinical or	Symptomatic; altered GI	Severely altered GI function;	Life-threatening consequences;	Death
	diagnostic observations only;	function	TPN or hospitalization indicated;	urgent intervention indicated
	intervention not indicated		elective operative intervention
			indicated
Definition: A disorder characterized by an abnormal communication between the rectum and another organ or anatomic site.
Rectal hemorrhage	Mild; intervention not indicated	Moderate symptoms; medical	Transfusion, radiologic,	Life-threatening consequences;	Death
		intervention or minor	endoscopic or elective	urgent intervention indicated
		cauterization indicated	operative intervention indicated
Definition: A disorder characterized by bleeding from the rectal wall and discharge from the anus.
Rectal mucositis	Asymptomatic or mild	Symptomatic; medical	Severe symptoms; limiting self	Life-threatening consequences;	Death
	symptoms; intervention not	intervention indicated; limiting	care ADL	urgent operative intervention
	indicated	instrumental ADL		indicated
Definition: A disorder characterized by inflammation of the mucous membrane of the rectum.
Rectal necrosis	–	–	Tube feeding or TPN indicated;	Life-threatening consequences;	Death
			radiologic, endoscopic, or	urgent operative intervention
			operative intervention indicated	indicated
Definition: A disorder characterized by a necrotic process occurring in the rectal wall.
Rectal obstruction	Asymptomatic; clinical or	Symptomatic; altered GI	Hospitalization indicated;	Life-threatening consequences;	Death
	diagnostic observations only;	function; limiting instrumental	elective operative intervention	urgent operative intervention
	intervention not indicated	ADL	indicated; limiting self care ADL;	indicated
			disabling
Definition: A disorder characterized by blockage of the normal flow of the intestinal contents in the rectum.
Rectal pain	Mild pain	Moderate pain; limiting	Severe pain; limiting self care	–	–
		instrumental ADL	ADL
Definition: A disorder characterized by a sensation of marked discomfort in the rectal region.
Rectal perforation	–	Symptomatic; medical	Severe symptoms; elective	Life-threatening consequences;	Death
		intervention indicated	operative intervention indicated	urgent operative intervention
				indicated
Definition: A disorder characterized by a rupture in the rectal wall.
Rectal stenosis	Asymptomatic; clinical or	Symptomatic; altered GI	Severely altered GI function;	Life-threatening consequences;	Death
	diagnostic observations only;	function	tube feeding or hospitalization	urgent operative intervention
	intervention not indicated		indicated; elective operative	indicated
			intervention indicated
Definition: A disorder characterized by a narrowing of the lumen of the rectum.
Rectal ulcer	Asymptomatic; clinical or	Symptomatic; altered GI	Severely altered GI function;	Life-threatening consequences;	Death
	diagnostic observations only;	function (e.g. altered dietary	TPN indicated; elective	urgent operative intervention
	intervention not indicated	habits, vomiting, diarrhea)	operative or endoscopic	indicated
			intervention indicated; disabling
Definition: A disorder characterized by a circumscribed, inflammatory and necrotic erosive lesion on the mucosal surface of the rectum.
Gastrointestinal disorders -	Asymptomatic or mild	Moderate; minimal, local or	Severe or medically significant	Life-threatening consequences;	Death
Other, specify	symptoms: clinical or diagnostic	noninvasive intervention	but not immediately life-	urgent intervention indicated
	observations only; intervention	indicated; limiting age-	threatening; hospitalization or
	not indicated	appropriate instrumental ADL	prolongation of existing
			hospitalization indicated;
			disabling; limiting self care ADL

Statistical Analysis

The unpaired t test was used to compare the averages of the two groups, while Fisher's exact test was used to compare the proportions. Kaplan-Meier curves were obtained for survival and toxicity rates, and the log-rank test was used to compare them. A P-value <0.05 was deemed statistically significant. Statistical analysis was performed with PASW Statistics 18 software (SPSS, Inc., Chicago, IL, USA).

RESULTS

Data for a total of 362 patients, all of whom had been treated for T1-4N0M0 adenocarcinoma of the prostate between 1998 and 2006, were collected from five representative institutions in Osaka, Japan. Postoperative cases were not included. None of the patients had been irradiated to the elective lymph node region, and all had been treated with the classical 4-field technique using 70 Gy in 35 fractions with gantry angles of 0°, 90°, 180° and 270°.

The median and mean ages of the patients were both 70 years (range, 49–82). The median follow-up period was 4.5 years (range, 1.0–11.6), with a minimum of 1 year. The actuarial 5-year overall and prostate cancer-specific survival rates were 93% and 96%, respectively (Fig. 1).

Fig. 1.

(a) The 5-year overall survival rate was 93%. (b) The 5-year prostate cancer-specific survival rate was 96%.

Neoadjuvant hormone therapy had been administered to 328 patients (91%), 35 of whom (11 %) had been considered hormone-refractory at the time of radiotherapy. Adjuvant hormone therapy had been administered to 276 of the total of 362 patients (76%), and 179 of them (65%) had already discontinued the therapy at the time of this survey. The median durations of neoadjuvant and adjuvant hormone therapy were 8 months (range, 1–150) and 24 months (range, 1–129), respectively.

2D simulation was performed for 127 patients, all of whom had been treated between 1998 and 2003. The other 235 had been treated using 3D simulation with a CT-simulator between 1998 and 2006. Of the five institutions, three had a 1 cm-width multileaf collimator (MLC), one a 2-cm MLC and one a 1-cm MLC until 2006, which was then replaced with a 0.5-cm MLC. The energy of the anterior-posterior beam was 10 MV at four institutions, and 20 MV at one. The energy of the lateral beams was 10 MV at three institutions, and 18 MV and 20 MV at one each. A field-shrinking technique was used for almost all patients (357 of 362, or 99%), once for 340 (94%) and twice for the other 17 (5%). The first shrinking was performed at 60 Gy for 219 patients (61%), at 50 Gy for 76 (21%), at 40 Gy for 42 (12%), and at other doses for 20 patients (6%). The second shrinking was performed at 60 Gy for 12 patients (71%), at 66 Gy for 4 (24%), and at 50 Gy for 1 (6%).

For the original irradiation field, three institutions defined the clinical target volume (CTV) as the prostate plus the whole seminal vesicle (SV), and the other two institutions as the prostate plus a part of SV. As for the shrinking field, three institutions defined the CTV as the prostate plus a part of SV, and the other two as the prostate only. In the original field, the median margin (distance from the CTV to the block edge) was 1.5 cm (range, 1.5–3.0) except for in the posterior (rectal) direction, where it was 1.3 cm (range, 1.0–2.0). As for the shrinking field, the median margin was 1.3 cm (range, 1.0–2.0) except for in the posterior direction, where it was 0.8 cm (range, 0.6–1.5). In 2D simulation, a Foley catheter was placed and contrast medium was administered into the bladder and rectum for visualization, trying to keep the definition of CTV and field margin described as above as much as possible. Retrograde urethrography was not performed routinely.

The use of the CT-simulator significantly reduced the irradiation field size compared to that used for 2D simulation (Table 3). The mean ± standard deviation (SD) of the distance between block edges in the right–left (RL) direction was 10.8 ± 1.1 cm for 2D simulation compared to 8.4 ± 1.2 cm for 3D (CT) simulation (P < 0.001). The corresponding values in the superior–inferior (SI) direction were 10.2 ± 1.0 cm and 8.2 ± 1.0 cm (P < 0.001), and in the anterior–posterior (AP) direction 8.8 ± 0.9 cm and 7.7 ± 1.0 cm (P < 0.001).

Table 3.

Comparison of 2D simulation and 3D (CT) simulation

	2D (n = 127)	3D (n = 235)	P
Median follow–up period (range) (year)	5.9 (1.1–11.6)	4.0 (1.0–8.0)	<0.001
Hormone therapy			<0.001
None	1 (1%)	30 (13%)
Neoadjuvant only	7 (6%)	43 (18%)
Adjuvant only	1 (1%)	1 (0%)
Both neoadjuvant and adjuvant	118 (93%)	161 (69%)
Multileaf collimator width			<0.001
0.5 cm	0 (0%)	5 (2%)
1.0 cm	127 (100%)	155 (66%)
2.0 cm	0 (0%)	75 (32%)
Portal filed size (cm)a
Right-left (RL)	10.8 ± 1.1	8.4 ± 1.2	<0.001
Superior-inferior (SI)	10.2 ± 1.0	8.2 ± 1.0	<0.001
Anterior-posterior (AP)	8.8 ± 0.9	7.7 ± 1.0	<0.001
Grade 1–3 late gastrointesitinal toxicity rate (%)			0.083
at 2 years	35	27
at 3 years	38	31
at 5 years	41	32
Grade 2–3 late gastrointesitinal toxicity rate (%)			<0.001
at 2 years	21	9
at 3 years	23	9
at 5 years	23	9
Grade 1–3 late rectal bleeding rate (%)			0.015
at 2 years	33	23
at 3 years	38	26
at 5 years	38	28
Grade 2–3 late rectal bleeding rate (%)			<0.001
at 2 years	21	7
at 3 years	23	7
at 5 years	23	7

aMean ± standard deviation.

Grade: Common Terminology Criteria for Adverse Events (CTCAE) Version 4.0.

Late gastrointestinal toxicity included late rectal bleeding.

Findings for toxicity are shown in Table 4. The maximum CTCAE Version 4.0 Grade toxicity was observed in the form of late genitourinary (GU) toxicity, late gastrointestinal (GI) toxicity including rectal bleeding, and late rectal bleeding alone. No Grade 4 or 5 late toxicity was observed; 5 patients (1%) suffered Grade 3 GU late toxicity and 10 (3%) Grade 3 GI late toxicity, all of which consisted of rectal bleeding; 14 patients (4%) suffered Grade 2 GU late toxicity and 35 (10%) Grade 2 late GI toxicity, 32 (9%) of which consisted of rectal bleeding. The actuarial 2-, 3-, and 5-year Grade 1–3 GU late toxicity rates were 13%, 17% and 23%, respectively, and the corresponding figures for Grade 2–3 were 2%, 4% and 6% (Fig. 2). The 2-, 3-, and 5-year Grade 1–3 GI late toxicity rates were 30%, 33% and 36%, respectively, and the corresponding figures for Grade 2–3 were 13%, 14% and 14% (Fig. 3). The 2-, 3-, and 5-year Grade 1–3 late rectal bleeding rates were 26%, 30% and 31%, respectively, and the corresponding figures for Grade 2–3 were 12%, 13% and 13%.

Table 4.

Grade of late toxicity

	Late genitourinary toxicity		Late gastrointestinal toxicity		Late rectal bleeding
	n	%	n	%	n	%
Grade 3	5	1	10	3	10	3
Grade 2	14	4	35	10	32	9
Grade 1	59	16	79	22	66	18
Grade 0	281	78	235	65	251	69
Missing	3	1	3	1	3	1
Total	362	100	362	101	362	100

Grade: Common Terminology Criteria for Adverse Events (CTCAE) Version 4.0.

Late gastrointestinal toxicity included late rectal bleeding.

Fig. 2.

Grade 2–3 late genitourinary toxicity. The 2-, 3- and 5-year toxicity rates were 2%, 4% and 6%, respectively.

Fig. 3.

Grade 2–3 late gastrointestinal toxicity. The 2-, 3- and 5-year toxicity rates were 13%, 14% and 14%, respectively.

When the patients were divided into a 2D- and a 3D-simulation group, the respective 2-, 3- and 5-year Grade 1–3 GI toxicity rates were 35%, 38% and 41% for 2D, and 27%, 31% and 32% for 3D (P = 0.083). The corresponding figures for Grade 2–3 were 21%, 23% and 23% for 2D, and 9%, 9% and 9% for 3D (P < 0.001) (Table 3). Similarly, the respective 2-, 3- and 5-year Grade 1–3 rectal bleeding rates were 33%, 38% and 38% for 2D, and 23%, 26% and 28% for 3D (P = 0.015), and the corresponding figures for Grade 2–3 were 21%, 23% and 23% for 2D, and 7%, 7% and 7% for 3D (P < 0.001) (Fig. 4).

Fig. 4.

Grade 2–3 late rectal bleeding. The 2-, 3- and 5-year occurrence rates were 21%, 23% and 23% for 2D simulation, and 7%, 7% and 7% for 3D simulation, respectively (P < 0.001).

DISCUSSION

To describe and analyse late toxicity is of the utmost importance for the use of radiation therapy for the treatment of prostate cancer. A number of publications have dealt with late toxicity in prostate radiotherapy [2-13]. However, most of these studies examined mixed populations with respect to prescribed dose, dose fractionation, or beam arrangements (for example, number of beam ports and their gantry angles). Therefore, the quantity of pure data for the effect of portal field size on late toxicity has been insufficient.

In Japan, prostate cancer was not considered to be a commonly occurring cancer until around 2000. Moreover, radical prostatectomy was preferred to radiotherapy by most urologists until that time [14]. However, the rate of prostate cancer incidence has been rapidly increasing recently [15], and at the same time, definitive radiotherapy has become the prevailing treatment mode. For these reasons, quite a few institutions did not have much experience with definitive radiotherapy for prostate cancer around 2000, when the subjects of our study were treated (1998–2006). Five representative institutions in the Osaka area, where radiation oncologists who had been trained at Osaka University were employed, participated in this study. These oncologists principally followed the same procedure as the one used at Osaka University, that is, the classical 4-field technique using anterior–posterior and lateral beams with 70 Gy in 35 fractions regardless of T-stage, Gleason Score or pretreatment prostate-specific antigen level. In fact, we found that 378 of all 436 patients (87%) enrolled in the previous survey study of ours had been treated with the same dose-fractionation of 70 Gy in 35 fractions. In view of this finding, we decided to embark upon this second survey to investigate solely the relationship between portal field size and late toxicity for a uniform setting of dose and beam arrangements.

To the best of our knowledge, our study cohort is one of the largest series treated with a uniform dose-fractionation and irradiation technique (classical 4-field technique). Moreover, all the institutions changed their simulation method from simple X-ray film-based simulation (2D) to CT simulation (3D) by the end of data acquisition for this study, which enabled us to compare the field size of 3D and 2D simulation. The results were very clear and easy to understand: 3D simulation reduced the field size significantly, as well as the rate of GI late toxicity, especially rectal bleeding. The reason for this improvement is deemed to be simply that with the smaller irradiation field the rectum could be largely avoided. There might be a speculation that not only field sizes but also the width of the MLC made an impact on the toxicities. We analysed the influence of the width of the MLC, but no statistically significant impact was detected (data not shown), we think because the number of patients treated with other than 1 cm-width MLC was too small compared to the 1 cm-width group. This issue should be addressed with other cohorts in other studies.

A strength of this study may well be that the surveyors were all physicians: no non-physicians participated. Moreover, they were mostly the same physicians that had treated the patients who were the subject of this paper. This makes a high degree of accuracy likely for the data collection, although it should be noted that this study was of a retrospective nature. On the other hand, a variation was observed in the incidence rate of Grade 1 toxicity among the five institutions as follows; 14–33% for 5-year Grade 1 GU late toxicity, 17–36% for GI and 13–30% for rectal bleeding. This variation might indicate that the incidence rate of Grade 1 depended on and was influenced by the physicians who followed up patients, especially in a retrospective analysis; therefore, the significance of the figures presented as Grade 1 toxicity should be considered as relatively low.

The rate of Grade 2–3 late toxicity detected in our study was similar to, or slightly higher than, the findings of other studies in the literature. Dearnaley et al. [6] reported that, in their randomized controlled trial in which all patients were treated with 64 Gy, radiation-induced Grade 2 or higher proctitis and bleeding occurred in 5% in the conformal group compared to 15% in the conventional group (P = 0.01). They found no difference between groups in bladder function after treatment (20 vs. 23% for Grade 2 or more, P = 0.61). It should be noted, however, that the toxicity scales used for their study were the Radiation Therapy Oncology Group (RTOG) criteria [16]. Morris et al. [7] conducted an evidence-based review of 3-dimensional conformal radiotherapy (3D-CRT) as part of an American Society for Radiation Oncology (ASTRO) outcomes initiative. In the Task Force Conclusion, they stated that 3D-CRT reduces late morbidity, particularly GI late morbidity, with the dose to the rectum limited. No benefits in terms of GU symptoms or sexual function were observed. Their conclusion thus shows good agreement with ours. Zelefsky et al., in their reports of the long-term results for 3D-CRT [8] and intensity-modulated radiotherapy (IMRT) [9] noted that, with 3D-CRT, the 5-year actuarial likelihood of Grade 2 and 3 late GI toxicities was 11% and 0.75%, respectively, while the corresponding findings for GU were 10% and 3%. With IMRT, the 10-year actuarial likelihood of Grade 2 and 3 late GI toxicities was 2% and 1%, respectively, while the corresponding findings for GU were 11% and 5%. The shapes of their actuarial toxicity curves resembled those of ours. That is, the GI toxicity curve reached a plateau at 2 or 3 years after radiotherapy, while the GU toxicity curve gradually rose until 10 years or more after radiotherapy. However, none of these studies provided detailed information on portal field size or its relation to late toxicity.

Dearnaley et al. had addressed this issue by a prospective randomized trial comparing 1.0 and 1.5 cm margins, arriving at the conclusion that the larger margin had been associated with the significantly higher incidence of toxicities [2]. However, their study had included only 126 patients, who had been assigned to 2 × 2 arms (64 Gy and 74 Gy groups, and, 1.0 and 1.5 cm margin groups). Moreover, their treatment planning had included two phases comprising a 3-field (anterior and left/right lateral or posterior oblique fields) phase and a 6-field (left and right, anterior/posterior oblique and lateral fields) phase. Although those patients had been randomly assigned, such critical heterogeneity of the cohort in total dose (64 Gy and 74 Gy) and treatment planning approach (3-field and 6-field) might make the interpretation complicated in terms of reproducibility. We considered that our current study could still add information and complement the conclusion drawn by Dearnaley et al., because it included a significantly larger number of patients (362 patients) and the treatment was in a more homogeneous manner (all with 70 Gy by 4-field) in spite of its weakness as a retrospective study.

The main criticism of our study might be that the kind of data on which it is based is so classical that no direct clinical indicators such as V40Gy or V65Gy of the rectum, could be provided as dose-volume constraints for modern 3D treatment planning for 3D-CRT or IMRT. However, the authors believe that the data presented here are still meaningful in terms of (i) describing a certain era of Japanese standard practice, (ii) providing radiation oncologists and treatment planners with a valuable reference because of the clear correspondence between a given portal field size (as a final block-to-block distance that would be relevant even for the most up-to-date irradiation technology) and a given rate of late toxicity, and (iii) providing suggestions for newly emerging irradiation technique in terms of a tolerance level that should not be exceeded, as detailed next.

CONCLUSION

In conclusion, we investigated late toxicity associated with EBRT for prostate cancer under conditions of a uniform setting of classical 4-field 70 Gy in 35 fractions. The use of CT simulation and the resultant reduction in the portal field size were significantly associated with diminished GI late toxicity, especially with less rectal bleeding. Typically, the field size was significantly reduced from 10.8 × 10.2 × 8.8 cm (2D simulation) to 8.4 × 8.2 × 7.7 cm (3D simulation), and at the same time, the rate of Grade 2–3 late rectal bleeding was significantly reduced from 23% to 7%. In view of the high overall and cause-specific survival rates observed in our study, any novel innovative radiotherapy should not exceed a late toxicity level of 7% for Grade 2–3 rectal bleeding in order to improve the quality of life of the patients or at least keep it the same as with “classical radiotherapy”.

FUNDING

This work was supported by Japan Society for the Promotion of Science (JSPS) KAKENHI (21791192).