LASSO NTCP predictorsforHNSCC and NPCLee et al.

LASSO NTCP predictors for the incidence of xerostomia inpatients with head and neck squamous cell carcinomaand nasopharyngeal carcinoma

Tsair-Fwu Lee1*, Ming-Hsiang Liou2,3, Yu-Jie Huang4,Pei-Ju Chao1,4, Hui-Min Ting2,4, Hsiao-Yi Lee2,Fu-Min Fang4

Supplementary information

Appendix1. Candidate predictive factors initiallyfor HNSCC patients in the xerostomia dataset
No. / Description / Range or Classification / (3m) Median or frequency / (3m)
correlation / (12M) Median or frequency / (12m) correlation
1 / Dmean-c / 4.9-38.4 / 23.2 / 0.654 / 22.5 / 1.422
2 / Dmean-i / 12.2-70 / 31.7 / 0.226 / 26.6 / 0.321
3 / Age / 34-74 / 53 / 0.193 / 51.9 / 0.151
4 / Gender / 0, 1# / 6, 80 / 1.35 / 3, 45 / 18.959
5 / Education / 0, 1, 2, 3 / 8, 17, 56, 5 / $I / 2, 4, 38, 4 / $VI
6 / Marriage / 0, 1* / 14, 72 / 1.184 / 7, 41 / 0.291
7 / Smoking / 0, 1* / 12, 74 / -0.652 / 10, 38 / 2.297
8 / Alcohol abuse / 0, 1* / 17, 69 / -0.223 / 11, 37 / 0.975
9 / T stage / 0, 1, 2, 3, / 9, 29, 7, 41 / $III / 4, 15, 4, 25 / $VIII
10 / Node classification / 0, 1 / 3, 83 / 0.811 / 0, 48 / -7.573
11 / Chemotherapy / 0, 1* / 20, 66 / 0.449 / 13, 35 / 0.061
12 / Baseline xerostomia / 0, 1 / 28, 58 / 0.262 / 17, 31 / 0.7
13 / Family history / 0, 1* / 65, 21 / -0.077 / 39, 9 / -1.633
14 / Financial status / 0, 1, 2, 3 / 23, 46, 13, 4 / $IV / 9, 26, 10, 3 / $IX
15 / Tumor site / 1, 2, 3, 4, 5 / 10, 13, 30, 31, 2 / $V / 6, 5, 12, 23, 2 / $X
16 / Surgery / 0, 1* / 42, 44 / -0.232 / 20, 28 / -0.43
17 / SIB or SQM / 0, 1 / 30, 56 / 0.642 / 18, 30 / 0.031

Abbreviation: Dmean-c: the mean dose given to the contralateral parotid gland;Dmean-i: the mean dose given to the ipsilateral parotid gland; SIB: simultaneous integrated boost; SQM: sequential mode; HNSCC: head and neck squamous cell carcinoma; NPC: nasopharyngeal carcinoma; LASSO: least absolute shrinkage and selection operator;XER: Xerostomia; XER3m or 12m:patient-reported moderate- to-severe xerostomia after 3- or 12-month;correlation was evaluated by univariate logistic regression;

# 0 = Female, 1 = Male;* 0 = No, 1 = Yes;Baseline xerostomia:xerostomia before RT，0 = No, 1 = a little;Education:0 = elementary, 1 = junior, 2 = senior, 3 = university;AJCC stage:1 = stage 1, 2 = stage 2, 3 = stage 3, 4 = stage 4;

T stage:0 = T0 and T1, 1 = T2(T2a, T2b), 2 = T3, 3 = T4;Node classification:0 = N0, 1 = N1~ N3(N3a, N3b);

Financial status:0 = under 16.7(f0), 1 = 16.7-33.3(f1), 2 = 33.3-66.7(f2), 3 = more than 66.7(f3); (unit:Thousand dollars);

Tumor site:1 =Larynx, 2 = Hypopharynx, 3 = Oropharynx, 4 = Oral cavity, 5 = Other;

$I:E0 = 0, E1 = 1.139, E2 = 1.786, E3 = 1.688;$II:stage1 = 0, stage 2 = -0.521, stage 3 = -20.073, stage 4 =0.878;$III:T0 = 0, T1 = -0.924, T2 = -20.088, T3 = -1.327;$IV:f0 = 0, f1 = -0.122, f2 = 0.554, f3 =0.393;$V:S1 = 0, S2 =0.424, S3 = 0.999, S4 = 1.041, S5 =-18.676; $VI:E0 = 0, E1 = -17.593, E2 = 0.632, E3 = -0.359; $VII:stage1 = 0, stage 2 = -1.449, stage 3 = 2.108, stage 4 = -0.664; $VIII:T0 = 0, T1 =-2.894, T2 =-23.36, T3 = -1.137; $IX:f0 = 0, f1 = 0.858, f2 = -0.79, f3 = 0.629; $X:S1 = 0, S2 =21.566, S3 = 21.658, S4 = 21.565, S5 = 1.549;

Appendix2.Candidate predictive factors initiallyfor NPC patients in the xerostomia dataset
No. / Description / Range or Classification / (3m) Median or frequency / (3m)
correlation / (12M) Median or frequency / (12m) correlation
1 / Dmean-c / 18.1-68.3 / 38.7 / 0.097 / 38.7 / 0.529
2. / Dmean-i / 20.7-69.3 / 41.3 / 0.058 / 41.6 / 0.505
3 / Age / 26-71 / 49 / 0.155 / 47.7 / 0.106
4 / Gender / 0, 1# / 18, 61 / -0.638 / 14, 44 / 0.02
5 / Education / 0, 1, 2, 3 / 2, 10, 41, 26 / $I / 0, 7, 29, 22 / $V
6 / Marriage / 0, 1* / 13, 66 / -0.296 / 9, 49 / 0.19
7 / Smoking / 0, 1* / 42, 37 / -0.402 / 34, 24 / 0.204
8 / Alcohol abuse / 0, 1* / 53, 26 / -0.138 / 39, 19 / -0.036
9 / T stage / 1, 2, 3, 4 / 30, 31, 6, 12 / $III / 21, 21, 5, 11 / $VII
10 / Node classification / 0, 1 / 16, 63 / 0.756 / 11, 47 / -0.142
11 / Chemotherapy / 0, 1* / 3, 76 / -0.225 / 1, 57 / -0.205
12 / Baseline xerostomia / 0, 1 / 41, 38 / 0.684 / 35, 23 / -0.177
13 / Family history / 0, 1* / 45, 34 / 0.218 / 34, 24 / -0.041
14 / Financial status / 0, 1, 2, 3 / 35, 37, 6, 1 / $IV / 28, 24, 6, 0 / $VIII
15 / SIB or SQM / 0, 1 / 49, 30 / -0.055 / 41, 17 / -0.268

Abbreviation: Dmean-c: the mean dose given to the contralateral parotid gland;Dmean-i: the mean dose given to the ipsilateral parotid gland; SIB: simultaneous integrated boost; SQM: sequential mode; HNSCC: head and neck squamous cell carcinoma; NPC: nasopharyngeal carcinoma; LASSO: least absolute shrinkage and selection operator;XER: Xerostomia; XER3m or 12m:patient-reported moderate- to-severe xerostomia after 3- or 12-month;correlation was evaluated by univariate logistic regression;

# 0 = Female, 1 = Male;* 0 = No, 1 = Yes;Education:0 = elementary, 1 = junior, 2 = senior, 3 = university;T stage:0 = T0 and T1, 1 = T2(T2a, T2b), 2 = T3, 3 = T4;Node classification:0 = N0, 1 = N1~N3(N3a, N3b);Baseline xerostomia:xerostomia before RT，0 = No, 1 = a little;Financial status:0 = under 16.7(f0), 1 = 16.7-33.3(f1), 2 = 33.3-66.7(f2), 3 = more than 66.7(f3); (unit:Thousand dollars); $I：E0 = 0, E1 =-21.644, E2 =-20.744, E3= -20.894; $II:stage1 = 0, stage 2 = -1.653, stage 3 = -0.15, stage 4 = -0.408; $III：T1 = 0, T2 = -0.528, T3 = -0.437, T4 =-0.89; $IV:f0 = 0, f1 = -0.535, f2 = 20.483, f3 = -21.909; $V：E0 = 0, E1 =0, E2 = 1.597, E3 = 2.676; $VI:stage1 = 0,stage 2 = -0.245, stage 3 = -0.196, stage 4 = -0.16; $VII:T1 = 0, T2 = 0.127, T3 = 0.402, T4 = 0.142; $VIII:f1 = 0, f2 = -0.127, f3 = -0.455;

1