Introduction
Weight loss (WL) is often observed during radiotherapy for head and neck squamous cell carcinoma (HNSCC). WL is mainly induced by radiation mucositis of the oral cavity and pharynx. Although the effects of WL before radiotherapy have consistently been reported to be associated with decreased overall survival, the effects of WL during radiotherapy are controversial [1–3].
WL during radiotherapy occurs even in modern precise radiotherapy, such as intensity-modulated radiation therapy (IMRT), including volumetric modulated arc therapy (VMAT), and leads to anatomical changes affecting dose distribution [4]. Although some studies suggest that routine replanning is not necessary during IMRT, it has a significant benefit in appropriately selected patients [5, 6].
Some studies have demonstrated the association between planning target volume (PTV) and WL during radiotherapy [7, 8]. However, there has been no detailed assessment of the relationship between prescribed doses, the volume of PTV, and PTV sites in the neck. Some institutions (including ours) use the lymph node regions adjacent to the primary tumor and/or metastatic lymph nodes as intermediate-risk PTV (IRPTV) and irradiate this area at higher doses than other prophylactic neck regions [low-risk PTV (LRPTV)] [9], despite the absence of gross tumors; however, the necessity of IRPTV remains unclear [10]. Therefore, this study aimed to evaluate the relationship between doses/volume/sites of the PTV and WL during IMRT treatment for HNSCC.
Materials and methods
Participants
In total, 79 patients with HNSCC who were treated with IMRT between January 2011 and December 2020 at our institution were reviewed. This retrospective study was approved by the institutional review board of our institution.
IMRT was delivered using a 6-MV X-ray from a linear accelerator (Varian Medical Systems, Inc., Palo Alto, CA, USA). The gross tumor volumes plus 5–10 mm margins were defined as high-risk clinical target volume (HRCTV). Lymph node regions adjacent to the primary tumor and/or metastatic lymph nodes were defined as intermediate-risk clinical target volumes (IRCTV). Whereas the prophylactic regions of the neck were defined as low-risk clinical target volume (LRCTV). The planning target volumes were made by adding 5 mm margins to the HRCTV, IRCTV, and LRCTV (HRPTV, IRPTV, and LRPTV, respectively). The most common treatment plan was as follows: 70 Gy delivered in 35 fractions to the HRPTV, 60 Gy delivered in 35 fractions to the IRPTV, and 54 Gy delivered in 35 fractions to the LRPTV.
Concurrent chemo/biotherapy was administered to 65 patients (platinum, n = 59; cetuximab, n = 6). Platinum chemotherapy consisted of cisplatin 80 mg/m2 every three weeks. The cisplatin dosage was reduced or switched to carboplatin, considering the general condition of the patients. Cetuximab was initiated one week before radiotherapy at a loading dose of 400 mg/m2, followed by a weekly infusion of 250 mg/m2 or the duration of radiotherapy.
The patients were divided into two groups using a cut-off of 10% WL (= severe WL). Body weight was measured at the beginning and every week during IMRT treatment. Image guidance for the setup was performed before all fractions of the IMRT treatment. In addition, WL was compared between the start of IMRT and the time of the most minimal weight. When oral intake became difficult during IMRT treatment, percutaneous endoscopic gastrostomy feeding, nasogastric tube feeding gastrostomy, or intravenous hyperalimentation was performed.
Statistical analyses
Statistical analyses were performed using the JMP software (JMP version 14.3.0; SAS Institute, Cary, NC, USA). Descriptive statistics were generated for percentage WL, as well as patient-, disease-, and treatment-related factors. Because there were no established optimal cutoff values for each PTV volume for predicting WL, receiver operating characteristic (ROC) curve analysis was performed. Fisher’s exact test was performed to test the different risk factor groups against the likelihood of a 10% WL. Stepwise regression analysis (a combination of forward selection and backward elimination) with the minimum corrected Akaike Information Criterion (AICc) was performed to select the optimal factors. Finally, the optimal risk factors were used in logistic regression analysis.
Results
Between January 2011 and December 2020, 95 patients with HNSCC were treated with IMRT, including VMAT, at our institution. Of these, 16 patients treated with three-dimensional radiotherapy and IMRT (hybrid radiotherapy) were excluded from the study. Finally, we retrospectively evaluated the remaining 79 patients with HNSCC (nasopharynx, 24; oropharynx, 18; hypopharynx, 30; larynx, 7) treated with IMRT using SIB methods. The details of these characteristics are shown in Table 1.
|
Characteristic |
No. of patients |
% |
|
|
Age |
Median 64 years (25–92 years) |
||
|
< 65 years |
44 |
55.7 |
|
|
≥ 65 years |
35 |
44.3 |
|
|
Sex |
Male |
67 |
84.8 |
|
Female |
12 |
15.2 |
|
|
Primary tumor sites |
Nasopharynx |
24 |
30.4 |
|
Oropharynx |
18 |
22.8 |
|
|
Hypopharynx |
30 |
38.0 |
|
|
Larynx |
7 |
8.9 |
|
|
Pretreatment BMI |
Median 20.4 (13.3–30.1) |
||
|
< 20 |
34 |
43,0 |
|
|
≥ 20 |
45 |
57,0 |
|
|
PS |
0 |
58 |
73.4 |
|
≥1 |
21 |
26.6 |
|
|
TNM (UICC 7th) |
cStage |
||
|
< 3 |
16 |
20.3 |
|
|
≥ 3 |
63 |
79.7 |
|
|
cT |
|||
|
< 3 |
43 |
54.4 |
|
|
≥ 3 |
36 |
45.6 |
|
|
cN |
|||
|
0 |
26 |
32.9 |
|
|
≥ 1 |
53 |
67.1 |
|
|
Systemic therapy |
Yes |
65 |
82.3 |
|
No |
14 |
17.7 |
|
|
Radiation dose |
HRPTV |
Median 70 Gy (66–70 Gy) |
|
|
IRPTV |
Median 60 Gy (60–63 Gy) |
||
|
LRPTV |
Median 54 Gy (54–56 Gy) |
||
The median WL and percentage of WL during IMRT treatment were 5.1 kg (range, 0–13.3 kg) and 8.5% (range, 0–18.2%), respectively. In addition, 31 patients (39%) had severe WL (≥ 10%) and 16 patients (20%) received nutritional support (percutaneous endoscopy gastrostomy, 3; nasogastric tube, 9; intravenous hyperalimentation, 4) due to disturbance of oral intake in the late phase of IMRT treatment.
Incidence of severe WL according to PTV volumes receiving ≥ 70 Gy, ≥ 60 Gy, and ≥ 54 Gy
The areas under the ROC curves for total PTV volumes receiving ≥ 70 Gy (PTV70Gy), ≥ 60 Gy (PTV60Gy), and ≥ 54 Gy (PTV54Gy) were 0.53 (sensitivity, 84%; specificity, 27%), 0.56 (sensitivity, 81%; specificity, 37%), and 0.59 (sensitivity, 94%; specificity, 37%), respectively. For severe WL, PTV54Gy, PTV70Gy, and PTV60Gy volumes of 615 cm3, 90 cm3, and 344 cm3, respectively, correspond to the maximum sum of sensitivity and specificity.
The incidence of severe WL was 25.0% (6/24), 45.5% (25/55), 13.6% (3/22), 49.1% (28/57), 31.3% (5/16), and 41.3% (26/63) in patients with PTV70Gy < 90 cm3, PTV70Gy ≥90 cm3 (p = 0.13), PTV60Gy < 344 cm3, PTV60Gy ≥ 344 cm3 (p < 0.01), PTV54Gy < 615 cm3, and PTV54Gy ≥ 615 cm3 (p = 0.57), respectively (Tab. 2). In addition, the incidence of severe WL was 28.3% (13/46) and 54.6% (18/33) in patients with IRPTV/PTV60Gy < 0.78 and IRPTV/PTV60Gy ≥ 0.78 (p = 0.02), respectively.
|
Factors |
> 10% WL |
p |
|
|
HRPTV + IRPTV + LRPTV (≥ 54 Gy) |
< 615 |
31.3% (5/16) |
0.06 |
|
≥ 615 |
41.3% (26/63) |
||
|
HRPTV + IRPTV (≥60 Gy) |
< 344 |
13.6% (3/22) |
< 0.01 |
|
≥ 344 |
49.1% (28/57) |
||
|
HRPTV (≥ 70 Gy) |
< 90 |
25.0% (6/24) |
0.13 |
|
≥ 90 |
45.5% (25/55) |
||
Site of the lymphatic region receiving ≥ 60 Gy and incidence of severe WL
The incidence of severe WL was 57.1% (8/14) in patients with PTV60Gy of ipsilateral level II neck, 13.3% (2/15) in patients with PTV60Gy of no level II neck (p = 0.02), 55.0% (11/20) in patients with PTV60Gy of ipsilateral level III neck, 12.5% (2/16) in patients with PTV60Gy of no level III neck (p = 0.01), 65.2% (15/23) in patients with PTV60Gy of bilateral level V neck, 25.7% (9/35) in patients with PTV60Gy of no level V neck (p = 0.01), 48.6% (17/35) in patients with PTV60Gy of bilateral level VII neck, and 20.0% (4/20) in patients with PTV60Gy of level VII neck (p = 0.05, Tab. 3).
|
Factors |
PTV region of ≥ 70 Gy |
PTV region of ≥ 60 Gy |
PTV region of ≥ 54 Gy |
||||
|
Lymphatic region |
> 10% WL |
p |
> 10% WL |
p |
> 10% WL |
p |
|
|
Level I |
No (control) |
38.7% (29/75) |
– |
32.3% (10/31) |
– |
28.6% (4/14) |
– |
|
Ipsilateral |
0% (0/1) |
> 0.99 |
34.5% (10/29) |
> 0.99 |
36.0% (9/25) |
0.73 |
|
|
Bilateral |
66.7% (2/3) |
0.56 |
57.9% (11/19) |
0.09 |
45.0% (18/40) |
0.35 |
|
|
Level II |
No (control) |
34.6% (9/26) |
– |
13.3% (2/15) |
– |
– |
– |
|
Ipsilateral |
40.0% (10/25) |
0.78 |
57.1% (8/14) |
0.02 |
– |
– |
|
|
Bilateral |
42.9% (12/28) |
0.59 |
42.0% (21/50) |
0.04 |
39.2% (31/79) |
– |
|
|
Level III |
No (control) |
28.2% (13/34) |
– |
12.5% (2/16) |
– |
– |
– |
|
Ipsilateral |
37.8% (11/29) |
> 0.99 |
55.0% (11/20) |
0.01 |
– |
– |
|
|
Bilateral |
43.8% (7/16) |
0.76 |
41.9% (18/43) |
0.03 |
39.2% (31/79) |
– |
|
|
Level IV |
No (control) |
37.5% (24/64) |
– |
23.1% (6/26) |
– |
0% (0/4) |
– |
|
Ipsilateral |
55.6% (5/9) |
0.47 |
33.3% (7/21) |
0.56 |
– |
– |
|
|
Bilateral |
33.3% (2/6) |
> 0.99 |
44.1% (15/34) |
0.11 |
41.3% (31/75) |
0.15 |
|
|
Level V |
No (control) |
39.5% (30/76) |
– |
25.7% (9/35) |
– |
21.4% (3/14) |
– |
|
Ipsilateral |
33.3% (1/3) |
> 0.99 |
33.3% (7/21) |
0.56 |
30.8% (4/13) |
0.68 |
|
|
Bilateral |
– |
– |
65.2% (15/23) |
0.01 |
46.2% (24/52) |
0.13 |
|
|
Level VII |
No (control) |
36.8% (25/68) |
– |
20.0% (4/20) |
- |
33.3% (2/6) |
– |
|
Ipsilateral |
71.4% (5/7) |
0.11 |
41.7% (10/24) |
0.20 |
28.6% (2/7) |
> 0.99 |
|
|
Bilateral |
25.0% (1/4) |
> 0.99 |
48.6% (17/35) |
0.05 |
40.9% (27/66) |
> 0.99 |
|
Incidence of severe WL according to other factors
The incidence of severe WL was 52.4% (22/42) in patients with hypopharyngeal or laryngeal cancer and 24.3% (9/37) in those with nasopharyngeal or oropharyngeal cancer (p = 0.01, Tab. 4). In addition, sex (male vs. female) and systemic therapy (yes vs. no) were statistically significant factors (p < 0.01 and 0.04, respectively) (Tab. 4) for the incidence of severe WL.
|
Factors |
> 10% WL |
p |
|
|
Age |
< 65 years |
40.9% (18/44) |
0.82 |
|
≥ 65 years |
37.1% (13/35) |
||
|
Sex |
Male |
31.3% (21/67) |
< 0.01 |
|
Female |
83.3% (10/12) |
||
|
Primary tumor sites |
Nasopharynx/oropharynx |
52.4% (22/42) |
0.01 |
|
Hypopharynx/larynx |
24.3% (9/37) |
||
|
Pretreatment BMI |
< 20 |
44.1% (15/34) |
0.49 |
|
≥ 20 |
35.6% (16/45) |
||
|
PS |
0 |
41.4% (24/58) |
0.61 |
|
≥ 1 |
33.3% (7/21) |
||
|
cStage |
<3 |
31.3% (5/16) |
0.57 |
|
≥ 3 |
41.3% (26/63) |
||
|
cT |
< 3 |
39.5% (17/43) |
> 0.99 |
|
≥ 3 |
38.9% (14/36) |
||
|
cN |
< 2 |
37.1% (13/35) |
0.82 |
|
≥ 2 |
40.9% (18/44) |
||
|
Systemic therapy |
Yes |
44.6% (29/65) |
0.04 |
|
No |
14.3% (2/14) |
||
|
Dmean of parotid gland |
< 25.9 |
30.6% (11/36) |
0,17 |
|
≥ 25.9 |
46.5% (20/43) |
||
|
Dmean of oral cavity |
< 41.7 |
30.0% (12/40) |
0.11 |
|
≥ 41.7 |
48.7% (19/39) |
||
Stepwise selection and multivariate analysis
Stepwise regression analysis with a cut-off p-value of 0.10 was performed including the abovementioned statistically significant clinical and PTV factors. The results revealed that three factors, primary site, PTV60Gy volume, and systemic therapy were selected for evaluation. In the multivariate analysis, primary site [odds ratio (OR): 3.0; 95% confidence interval (CI): 1.0–8.5; p = 0.04) and PTV60Gy ≥ 344 cm3 (OR: 4.7; 95% CI: 1.0–24.4; p = 0.04, Tab. 5) were significant independent unfavorable factors for severe WL during IMRT treatment. Systemic therapy was not a significant unfavorable factor for severe WL (OR: 3.3; 95% CI: 0.6–19.3; p = 0.18, Tab. 5).
|
Factors |
OR (95% CI) |
p |
|
|
Primary site |
Hypopharynx/larynx |
3.0 (1.0–8.5) |
0.04 |
|
Nasopharynx/oropharynx |
|||
|
HRPTV + IRPTV (≥ 60 Gy) |
< 344 |
4.7 (1.0–24.4) |
0.04 |
|
≥ 344 |
|||
|
Systemic therapy |
No |
3.3 (0.6–19.3) |
0.18 |
|
Yes |
|||
PTV60Gy ≥ 344 cm3 was significantly correlated with radiation-induced mucositis ≥ Grade 3 of Common Terminology Criteria for Adverse Events version 4.0 (CTCAE v4.0) (p = 0.02); however, primary site was not correlated with radiation-induced mucositis ≥ Grade 3 (p = 0.62). In addition, the incidence of severe WL with mucositis ≥ Grade 3 vs. < Grade 3 was 57.9% (11/19) and 23.5% (4/17), respectively (p = 0.05).
Discussion
In our study, severe WL during IMRT treatment was associated with primary site and PTV60Gy volume. Sites of PTV60Gy were not the parameters that had a high correlation with the severe WL during IMRT treatment.
Mallick et al. investigated factors associated with WL during radiotherapy [8]. They mentioned that the total PTV (> 615 cm3) and PTV70Gy (> 235 cm3) were statistically significant factors for predicting WL during radiotherapy treatment. In contrast, our study suggested that a large PTV 60 Gy volume (≥ 344 cm3) was a statistically significant factor for severe WL during IMRT. At present, the necessity of large IRPTV (60 Gy) regions in IMRT for head and neck cancer is unclear [10]. Lee et al. mentioned that the lymph node regions adjacent to the primary tumor and/or metastatic lymph nodes could be considered as IRPTV [9]. In contrast, Hansen et al. stated that the 5 mm margin to the primary tumor and/or metastatic lymph nodes should be considered as IRPTV [11]. Although dose reduction in the LRPTV (54 Gy) and range reduction of IRPTV have been attempted in recent clinical trials and guidelines [11–14], a smaller IRPTV seemed to be preferable in terms of severe WL during IMRT. Furthermore, it may not be necessary to perform IRPTV since there has been no data that indicates IRPTV affects the efficacy of IMRT for head and neck cancer.
In addition, Langius et al. reported the impact of PTV regions (ipsilateral or bilateral vs. no) on WL during radiotherapy [7]. However, in their study, two different types of irradiation techniques (IMRT and three-dimensional conformal radiotherapy) were used, which also included patients who received postoperative radiotherapy. Some studies have reported that these factors are important for WL during radiotherapy treatment [8, 15]. In our study, these two factors were excluded. In addition, the PTV regions were divided into lymphatic regions and analyzed to accommodate the IMRT era. Although PTV60Gy regions (ipsilateral level II or III and bilateral level V or VII) were statistically significant factors in univariate analysis, these PTV60Gy regions were not parameters that had the highest correlation with severe WL during IMRT treatment in stepwise and multivariate analyses. Furthermore, these PTV60Gy regions were thought to be confounding factors, and the PTV60Gy volume was found to be more important than the PTV 60 Gy region. In our study, severe mucositis (≥ Grade 3 of the CTCAE v4.0) was correlated with PTV60Gy ≥ 344 cm3 and influenced severe WL during IMRT treatment. We believe that a large PTV60Gy volume increased mucositis and led to severe WL during IMRT. Although IRPTV was defined as 5 mm margins surrounding the lymph node region adjacent to the primary tumor and/or metastatic lymph nodes, this may have too large margins in terms of severe WL during IMRT.
In our study, the primary tumor site was the significant factor for severe WL during IMRT treatment. This has been reported to influence WL during radiotherapy in some studies [16, 17]. The primary tumor site was included in the abovementioned factor because it is part of the PTV60Gy regions. Because this factor was unchangeable in HNSCC treatment, this would be important as useful predictors for severe WL during IMRT treatment.
There were some limitations to our study owing to its retrospective nature. First, the sample size is small. Therefore, it was necessary to select factors for the multivariate analysis using a stepwise selection. However, because systemic therapy improved treatment outcomes even in elderly patients with HNSCC [18], systemic therapy should be combined with radiotherapy for HNSCC, even if it is a factor associated with severe WL. Second, although the use of nutritional support during IMRT treatment in our study was determined by each physician according to each case, the frequency was sufficiently low. This was important when considering the risk of true severe WL during IMRT treatment because some patients with head and neck cancer prefer to be treated on an outpatient basis. Because only a few reports have examined WL during radiotherapy treatment, we believe that our study showed an important finding in daily clinical practice. Furthermore, recently, dose calculation algorithms have been improved [19]. When dose calculation algorithm improves, it may also affect the dose distribution of IMRT plans. Therefore, updates will be needed regarding WL during IMRT treatment as dose calculation algorithm improves.
In conclusion, a large PTV60Gy (especially in level II or III neck regions) was associated with severe WL during IMRT treatment. Because one of the risk factors for severe WL during IMRT treatment was PTV60Gy, the range reduction of IRPTV seemed to be important in terms of severe WL.
Acknowledgements
None declared.
Conflict of interest
The authors declare that they have no conflict of interest.
Funding
This work did not receive any specific grant from founding agencies in the public, commercial, or not-for-profit sectors.