EPIC scores were recorded prior to and 20–36 months following IMRT in 445 patients. See Table 1 for patient characteristics. The baseline domains and subscales of definitive RT patients (n = 264) were compared RP patients (n = 50). RP patients had significantly higher sexual bother (median 81.25 vs. 50, p = 0.0184) and overall (median 53.23 vs. 30.77, p = 0.0212) scores. There were no other significant differences. Post-operative patients had significantly worse baseline urinary function, incontinence, and overall score (all p-value < 0.0001), but not bother or irritative scores compared to definitive RT patients. There were no other significant differences.
Table 1
| | | n (%) |
T stage* | 1 | | 128 (51) |
2 | | 81 (32) |
3 | | 44 (17) |
N stage* | 1 | | 12 (5) |
Gleason† | ≤ 6 | | 81 (21) |
| 7 | | 198 (51) |
| ≥ 8 | | 108 (28) |
Race | White | | 168 (70) |
Black | | 53 (22) |
Other | | 20 (8) |
Definitive RT | PSA pre-RT median (range), ng/ml | 7.8 (0.1–380) |
Nodal Dose (Gy) | 0 | 132 (47) |
45 | 18 (6) |
54 | 131 (47) |
Hormones†† | | 71 (46) |
Postoperative RT | PSA pre-RT median (range), ng/ml | 0.2 (0.0–25) |
Adjuvant | | 44 (28) |
Salvage | | 116 (73) |
Nodal Dose (Gy) | 0 | 5 (3) |
45 | 18 (11) |
54 | 144 (86) |
Hormones†† | | 22 (23) |
Abbreviations: T, tumor; N, node; PSA, prostate specific antigen; RT, radiotherapy; Gy, gray *pathologic staging for postoperative and clinical for definitive patients †from surgery specimen for postoperative and biopsy for definitive ††Immediately prior to, during, or after radiotherapy prior to follow-up |
Patients treated to 0, 45 and 54 Gy had a MID in overall urinary decline at frequencies of 30%, 26%, and 42%, respectively. MID declines were also found in urinary function (29%, 26%, 46%), bother (31%, 36%, 40%), irritative (24%, 36%, 30%), and incontinence (35%, 27%, 53%). Patients treated to 0, 45 and 54 Gy with definitive RT had a MID overall urinary decline in 31%, 20%, and 35%, respectively. MID declines were also found in urinary function (30%, 20%, 41%) bother (33%, 31%, 44%), irritative (25%, 27%, 23%), and incontinence (36%, 19%, 52%). Patients treated to 0, 45 and 54 Gy with post-operative RT had a MID overall urinary decline in 0%, 31%, and 49%, respectively. MID declines were also found in urinary function (0%, 31%, 50%), bother (0%, 41%, 45%), irritative (0%, 44%, 36%), and incontinence (0%, 35%, 54%). Frequency of MID decline was significantly greater in patients treated to 54 Gy for urinary function, incontinence, and overall (p = 0.0024, 0.0007, 0.0218, respectively). On subset analysis of definitive and postoperative patients, frequency of MID incontinence decline remained significant (p = 0.0148 and 0.0225, respectively). In all patients, MID declines were experienced in bowel function (30%, 32%, 33%), bother (34%, 27%, 38%), and overall (34%, 29%, 38%). See Table 2.
Table 2
Patients Experiencing a MID Decline by Nodal Dose
| | All Patients | Definitive Radiotherapy | Postoperative Radiotherapy |
Domain | Subscale/Summary Score | 0 Gy (n = 131) | 45 Gy (n = 33) | p-value* | 54 Gy (n = 264) | p-value* | 0 Gy (n = 126) | 45 Gy (n = 16) | p-value* | 54 Gy (n = 125) | p-value* | 0 Gy (n = 5) | 45 Gy (n = 16) | p-value* | 54 Gy (n = 139) | p-value* |
Urinary | Function | 28.9% | 25.8% | 0.7308 | 45.5% | 0.0024 | 30.2% | 20.0% | 0.5519 | 40.9% | 0.0893 | 0.0% | 31.3% | 0.2776 | 49.6% | 0.0591 |
Bother | 31.3% | 36.4% | 0.5781 | 39.8% | 0.1004 | 32.5% | 31.3% | 0.9173 | 34.4% | 0.7548 | 0.0% | 41.2% | 0.1348 | 44.6% | 0.0702 |
Irritative | 23.5% | 35.5% | 0.1765 | 29.8% | 0.2143 | 24.6% | 26.7% | 1.0000 | 23.0% | 0.7836 | 0.0% | 43.8% | 0.1235 | 35.7% | 0.1645 |
Incontinence | 35.1% | 27.3% | 0.3951 | 53.0% | 0.0007 | 36.4% | 18.8% | 0.1604 | 51.6% | 0.0148 | 0.0% | 35.3% | 0.2663 | 54.4% | 0.0225 |
Overall | 29.8% | 25.8% | 0.6656 | 42.2% | 0.0218 | 31.0% | 20.0% | 0.5510 | 34.5% | 0.5749 | 0.0% | 31.3% | 0.2776 | 48.8% | 0.0601 |
Bowel | Function | 29.8% | 32.3% | 0.7864 | 32.9% | 0.5424 | 31.0% | 46.7% | 0.2504 | 32.7% | 0.7814 | 0.0% | 18.8% | 0.5489 | 33.1% | 0.1774 |
Bother | 34.3% | 27.3% | 0.4398 | 38.0% | 0.4704 | 35.7% | 31.3% | 0.7276 | 40.3% | 0.4447 | 0.0% | 23.5% | 0.5352 | 36.0% | 0.1635 |
Overall | 33.9% | 29.0% | 0.6079 | 37.5% | 0.5000 | 35.3% | 46.7% | 0.3918 | 38.4% | 0.6334 | 0.0% | 12.5% | 1.0000 | 36.7% | 0.1609 |
Abbreviations: MID, minimally important difference; Gy, gray |
*compared to 0 Gy |
On MVA by absolute change, pretreatment score was significant for all domains and subscales. Postoperative status was significant for decline in urinary function, incontinence, and overall. PLNRT to 54 Gy was significant for decline in urinary function, incontinence, and overall. Year of RT was significant for all urinary categories. Age was significant for decline in bowel function. Year of RT was significant for decline in bowel bother and overall scores. Pre-treatment score was the only variable significant for decline of hormone function, bother and overall. See Table 3 for p-values of MVA using absolute score change or MID frequency (decline) and association with bowel and urinary domain and subscale score changes.
Table 3
| | MVA model | Age: 1-year increase | Pre-treatment Score: 1-point increase | Post op: Yes vs No | Node Dose: 45–51 vs 0 | Node Dose: 54 vs 0 | Year of RT: 1-year increase | Step and Shoot*: Yes vs No | EPIC follow-up: 1-year increase |
Urinary | Function | Absolute | 0.8453 | < 0.0001 | 0.0012 | 0.95 | 0.0334 | 0.0277 | 0.7452 | 0.898 |
MID | 0.9542 | < 0.0001 | 0.0267 | 0.9252 | 0.0003 | 0.0743 | 0.9917 | 0.6957 |
Bother | Absolute | 0.7467 | < 0.0001 | 0.4359 | 0.4653 | 0.0933 | 0.0107 | 0.6277 | 0.3632 |
MID | 0.8424 | < 0.0001 | 0.1349 | 0.5159 | 0.0015 | 0.0647 | 0.4219 | 0.2135 |
Irritative | Absolute | 0.8977 | < 0.0001 | 0.9474 | 0.6603 | 0.0959 | 0.0248 | 0.3324 | 0.8949 |
MID | 0.4652 | < 0.0001 | 0.3344 | 0.579 | 0.0101 | 0.4619 | 0.1579 | 0.175 |
Incontinence | Absolute | 0.7757 | < 0.0001 | 0.046 | 0.5881 | 0.0492 | 0.0468 | 0.5795 | 0.4617 |
MID | 0.9673 | < 0.0001 | 0.2164 | 0.6938 | < 0.0001 | 0.0525 | 0.8028 | 0.9439 |
Overall | Absolute | 0.9968 | < 0.0001 | 0.0182 | 0.7355 | 0.0367 | 0.0072 | 0.7849 | 0.8023 |
MID | 0.9797 | < 0.0001 | 0.0045 | 0.8059 | 0.0002 | 0.0077 | 0.7342 | 0.7126 |
Bowel | Function | Absolute | 0.0436 | < 0.0001 | 0.7653 | 0.871 | 0.3036 | 0.1117 | 0.7617 | 0.1575 |
MID | 0.0458 | < 0.0001 | 0.3622 | 0.7742 | 0.0083 | 0.2847 | 0.2647 | 0.2861 |
Bother | Absolute | 0.9799 | < 0.0001 | 0.6065 | 0.8133 | 0.1016 | 0.0329 | 0.9174 | 0.3272 |
MID | 0.147 | < 0.0001 | 0.3316 | 0.9823 | 0.0781 | 0.3937 | 0.5133 | 0.8148 |
Overall | Absolute | 0.2609 | < 0.0001 | 0.5722 | 0.985 | 0.1954 | 0.0447 | 0.982 | 0.1633 |
MID | 0.0912 | < 0.0001 | 0.2679 | 0.9026 | 0.063 | 0.3105 | 0.3115 | 0.1334 |
Hormone | Function | Absolute | 0.5588 | < 0.0001 | 0.892 | 0.3069 | 0.0802 | 0.1707 | NA | 0.7164 |
Bother | Absolute | 0.6627 | < 0.0001 | 0.0919 | 0.6831 | 0.9597 | 0.3493 | NA | 0.835 |
Overall | Absolute | 0.9004 | < 0.0001 | 0.4663 | 0.4817 | 0.172 | 0.2528 | NA | 0.6843 |
Abbreviations: MVA: multiavariate analysis; RT, radiotherapy; EPIC, Expanded Prostate Cancer Index Composite |
*Patients all received Intensity Modulated Radiotherapy via either 'Step and Shoot' or 'arc.' |
Amongst postoperative patients, those treated in the salvage setting had a greater decline in urinary function score (p = 0.0254) with a trend toward greater decline in urinary incontinence (p = 0.0564) and overall (p = 0.0714) scores.
DOSIMETRIC ANALYSIS
Genitourinary
No significant dosimetric relationships were found for urinary irritative change in definitive or postoperative patients. Urinary function change weakly negatively correlated with V20 in postop patients (cut-point for V20 of 100% with a PPV of 57%). No other significant relationships were found between dosimetric parameters and subscales postoperatively.
In definitive patients, V20, 30,40,50,60 and 70 all negatively correlated with urinary incontinence change. Isodose volume cut-points each showed a MID decline with positive predictive value (PPV) of > 44%: V70 of 9%, V60 of 16%, V50 of 31%, V40 of 50%, V30 of 71%, and V20 of 97%. A significant difference was detected between AUCs for these isodose volumes (p-value = 0.0132). The V20 cut point of 97% had the highest AUC (p-values all ≥ 0.13 by pairwise tests). The V20 moderately correlated (cut-point for V20 of 98% with a PPV of 46%) with urinary function change while V30, 40, 50, and 60 weakly correlated (cut-point of 16–71% with a PPV of 39–46%). A significant difference between AUCs was not detected for these isodose volumes (p-value = 0.1612). Urinary bother changes weakly correlated with V30 (cut-point of 72% with a PPV of 32%).
Urinary overall score change had significant correlations only in definitive patients (V20-50, cut-points of 31–98% with PPVs of 29–36%). A significant difference was detected between AUCs for these isodose volumes (p-value = 0.0359). The V20 cut point of 98% had the highest AUC and was significantly higher than V40, the lowest AUC. See Fig. 1 and Table 4.
Table 4
Urinary EPIC Score Changes by Bladder Dosimetry
Setting | Bladder Volume (v in %) at Dose (Gy) | Function* | p-value | Bother* | p-value | Irritative* | p-value | Incontinence* | p-value | Overall* | p-value |
All | v75 | 0.14685 | 0.0321 | 0.11531 | 0.071 | 0.16326 | 0.0175 | 0.0779 | 0.2218 | 0.16485 | 0.0159 |
v70 | -0.10592 | 0.1234 | -0.03528 | 0.5822 | 0.03756 | 0.5879 | -0.10734 | 0.0917 | -0.07512 | 0.2754 |
v60 | -0.19683 | 0.0039 | -0.07596 | 0.2355 | 0.00163 | 0.9813 | -0.17474 | 0.0057 | -0.1507 | 0.0278 |
v50 | -0.23562 | 0.0005 | -0.11447 | 0.0731 | -0.02415 | 0.7276 | -0.22393 | 0.0004 | -0.1927 | 0.0047 |
v40 | -0.2522 | 0.0002 | -0.14307 | 0.0247 | -0.04634 | 0.5037 | -0.24866 | < .0001 | -0.20991 | 0.002 |
v30 | -0.26136 | 0.0001 | -0.16277 | 0.0105 | -0.06992 | 0.3125 | -0.26108 | < .0001 | -0.22188 | 0.0011 |
v20 | -0.31993 | < .0001 | -0.15759 | 0.0132 | -0.08414 | 0.2238 | -0.28921 | < .0001 | -0.25188 | 0.0002 |
Definitive | v75 | -0.01253 | 0.8868 | 0.02128 | 0.7937 | 0.10927 | 0.2163 | -0.0571 | 0.4795 | 0.01627 | 0.8534 |
v70 | -0.12653 | 0.1485 | -0.03683 | 0.6507 | 0.05428 | 0.5403 | -0.16455 | 0.04 | -0.08334 | 0.3428 |
v60 | -0.20181 | 0.0201 | -0.0432 | 0.5953 | 0.0868 | 0.3267 | -0.22515 | 0.0046 | -0.12657 | 0.1484 |
v50 | -0.24937 | 0.0038 | -0.08582 | 0.2904 | 0.05672 | 0.5222 | -0.29016 | 0.0002 | -0.1721 | 0.0483 |
v40 | -0.27382 | 0.0014 | -0.14087 | 0.0814 | 0.000261 | 0.9977 | -0.32017 | < 0.0001 | -0.20358 | 0.019 |
v30 | -0.25999 | 0.0025 | -0.16684 | 0.0385 | -0.04132 | 0.6412 | -0.31637 | < 0.0001 | -0.21406 | 0.0135 |
v20 | -0.3035 | 0.0004 | -0.14372 | 0.0753 | -0.03981 | 0.6535 | -0.34752 | < 0.0001 | -0.22265 | 0.0101 |
Postop | v75 | 0.0611 | 0.589 | 0.08085 | 0.4447 | 0.09415 | 0.4043 | 0.03579 | 0.7356 | 0.08717 | 0.4402 |
v70 | 0.04405 | 0.6971 | 0.04537 | 0.6684 | 0.09856 | 0.3825 | 0.01614 | 0.879 | 0.0627 | 0.5793 |
v60 | -0.00753 | 0.947 | -0.0193 | 0.8555 | 0.02028 | 0.8578 | -0.02253 | 0.8317 | -0.00314 | 0.9779 |
v50 | -0.06703 | 0.5533 | -0.06481 | 0.5404 | -0.03041 | 0.7882 | -0.05509 | 0.6029 | -0.06658 | 0.5559 |
v40 | -0.08811 | 0.4353 | -0.04689 | 0.658 | -0.02808 | 0.8041 | -0.07886 | 0.456 | -0.07519 | 0.5059 |
v30 | -0.14584 | 0.1945 | -0.0474 | 0.6545 | -0.03281 | 0.7719 | -0.09463 | 0.3706 | -0.09968 | 0.3771 |
v20 | -0.25103 | 0.0235 | -0.09719 | 0.3577 | -0.0864 | 0.4443 | -0.1044 | 0.3229 | -0.19066 | 0.0882 |
*Spearman Rank Correlation |
Using 16 dosimetrically representative 54 Gy patients, planning to 45 Gy would have resulted in a significantly decreased bladder V20 (median 93.5 [75–100 Gy], vs 98.0 [90–100 Gy], p = 0.0086). Bladder V30 (median 65 [48–97 Gy], vs 75 [54–99 Gy], p = 0.0837), V40 (median 44 [28–71 Gy], vs 50 [35–72 Gy], p = 0.0770), and V50 (median 24.5 [13–50 Gy], vs 30.5 [22–53 Gy], p = 0.1057) were decreased without reaching significance. Bladder V60 and V70 were nearly identical (median values within 0.5).
Gastrointestinal
No significant correlations were found for bowel function, bother, or overall change. A significant, weakly negative correlation was found for small bowel V40 and V45 and bowel bother change, with respective cut-points at 6% (37% PPV) and 3% (34% PPV). A significant difference between AUCs was not detected for these isodose volumes (p-value = 0.1376).
Per dosimetrically representative sampling, planning to 45 Gy would have resulted in a significantly decreased small bowel V40 (median 3 [0–18 Gy], vs 11 [1–54 Gy], p = 0.0058), V45 (median 0 [0–3 Gy], vs 5 [0–36 Gy], p < 0.0001), and V50 (median 0 [0–0 Gy], vs 1 [0–7 Gy], p = 0.0007) Small Bowel V52 (median 0 [0–0 Gy], vs 0 [0–2 Gy], p = 0.1639) was decreased without reaching significance. There were no significant differences in rectal V20-75.