3 D Facial Imaging Project Tatiana Foroud Elizabeth

3 D Facial Imaging Project Tatiana Foroud Elizabeth Moore Richard Ward Leah Wetherill Li Shen Jeffrey Rogers Shiaofen Fang

Data Collection (3 d. MD only) (as of January 3, 2011) Site 3 D Images (% of target) DNA San Diego 188 (125%) 157 UCLA 52 (104%) 80 Atlanta 116 (93%) 54 Ukraine 35 (37%) 0 (no approval) Cape Town 224 (100%) 225* Totals 599 516 * Also obtained 218 DNA samples from the mothers and 52 DNA samples from the father. RNA also collected at this site for the subject

DNA n Saliva was obtained from US participants n n DNA has been extracted Will perform a GWAS at CIDR on 250 US participants Data to be analyzed primarily for hypothesis generation n Insufficient power to support a GWAS for FAS n

Pass Collaboration n Applied for and received approval to collect 3 D images at 1 month and 12 months n Sample of 1200 will have longitudinal assessment Training completed by Leah Wetherill and Mike Suttie earlier this month n Collection began last week n

Pass Collaboration

Goal 1: Improve understanding of the dysmorphic features in FAS n Focused on site with the largest number of 3 D images (South Africa) FAS vs. controls n Alcohol exposed vs. controls n n Goal is to delineate facial characteristics affected by alcohol exposure Requires larger numbers in each group n Novel approaches to analyze single subjects being developed by Peter Hammond n

Peter Hammond Green: no difference between control and heavy exposed Red: heavy alcohol exposed smaller than control Blue: heavy alcohol exposed larger than control • Illustrates the power of a large number of controls • Focus future data collection to ensure large control groups of similar age and ethnicity Peter Hammond

Goal 2: Enhance the capability of definitive diagnosis of fetal alcohol syndrome (FAS) and the broader spectrum of fetal alcohol spectrum disorders (FASD) at different stages of the lifespan n Comparison using data from South Africa n Each time point, age of sample is relatively homogenous n Time point 1 – average 5 years n Time point 2 – average 9 years ~90% of subjects seen at both time points n Data from 2 cameras n

South Africa Longitudinal Sample FAS PFAS Heavy exposure Control Total Visit 1 17 24 47 57 145 Visit 2 24 26 83 87 220 Both 17 23 46 57 143 Both* 14 21 41 52 128 CIFASD FAS vs. Heavy Prenatal Alcohol Exposure vs. Controls

Anthropometric Measurements 7 widths: MF: Minimal frontal (ft-ft) BZ: Bizygoma (zy-zy BT: Bitragal (t-t) BG: Bigonial (go-go) OC: Outer eye (ex-ex) IC: Inner eye (en-en) PF: Palpebral fiss (en-ex) 6 lengths: NL: Nasal (n-sn) NBL: Nasal bridge (n-prn) PL: Philtrum (sn-ls) LF: Lower face (sn-gn) TFH: Total face (n-gn) REL: Ear (sa-sba) 3 depths: RUF: Upper face (t-n) RMF: Mid face (t-sn) RLF: Lower face (t-gn)

Discrimination n Used discriminant analyses to build models with the anthropometric measures n Compared 2 models n FAS vs. Control; Alcohol exposed vs. Control n Used stepwise models to add/remove variables n Difficult to compare variables in models since some measures correlated n Tried principal components with mixed results

Compare Measurements n Comparing means of anthropometric measurements at 2 time points between group definitions Comparable to approach in Aim 1 n Less focus on model building for diagnosis, more focus on understanding biological differences n Easier comparison to animal models n Consider greater subdivision of categories n

Results Visit 1 (Ave Age 5) FAS vs PFAS vs HE vs group p PFAS HE Control HE control minimal frontal 0. 0001 0. 0208 0. 0001 0. 0049 bizygomatic 0. 0001 0. 0058 0. 0002 0. 0001 bitragal 0. 0004 0. 0051 0. 0005 0. 0001 bigonial 0. 0579 0. 0352 0. 0098 0. 0117 inner canthal outer canthal 0. 0001 0. 0009 palpebral fissure 0. 0023 0. 0001 0. 0003 0. 0004 upper facial depth 0. 0003 0. 0076 0. 0001 mid facial depth 0. 0001 0. 0037 0. 0001 lower facial depth 0. 0001 0. 0025 0. 0001 0. 017 0. 0199 nose length 0. 0112 0. 0198 0. 0018 0. 0503 nasal bridge length 0. 0006 0. 0026 0. 0001 0. 0089 philtrum length lower facial height total facial height 0. 0006 0. 002 0. 0001 0. 0303 ear length 0. 0001 0. 0016 0. 0002 Age 0. 0092 0. 0036 0. 0016

Results Visit 2 (Ave Age 9) FAS vs PFAS vs HE vs group p PFAS HE Control HE control minimal frontal 0. 0001 0. 052 0. 0008 0. 0001 0. 0086 bizygomatic 0. 0001 0. 0074 0. 0001 0. 0098 0. 0062 bitragal 0. 0001 0. 0261 0. 0001 0. 0987 0. 0119 bigonial 0. 0011 0. 0443 0. 0004 0. 0002 inner canthal outer canthal 0. 0001 0. 0458 0. 0001 0. 0086 palpebral fissure 0. 0001 0. 0158 0. 0001 0. 0011 0. 0534 upper facial depth 0. 0001 0. 034 0. 0001 0. 0201 0. 0037 mid facial depth 0. 0001 0. 0218 0. 0001 0. 0187 0. 0011 lower facial depth 0. 0001 0. 0057 0. 0001 0. 0277 0. 0004 nose length 0. 0001 0. 0132 0. 001 0. 0078 0. 0093 nasal bridge length 0. 0003 0. 0515 0. 0078 0. 0001 0. 0238 0. 0301 philtrum length lower facial height total facial height 0. 0001 0. 0264 0. 0002 0. 0001 0. 0177 ear length 0. 0001 Age 0. 0085 0. 0033 0. 0016

Longitudinal Change group p minimal frontal bizygomatic 0. 0001 bitragal bigonial inner canthal outer canthal palpebral fissure upper facial depth 0. 008 mid facial depth 0. 0101 lower facial depth 0. 015 nose length nasal bridge length philtrum length lower facial height total facial height ear length 0. 041 Age 0. 0085 FAS vs PFAS FAS vs HE 0. 0009 FAS vs Control 0. 0102 PFAS vs HE control 0. 0018 0. 0252 0. 027 0. 015 0. 0044 0. 0420 0. 0067 0. 0190 0. 0033 0. 059 0. 0016 HE vs control 0. 037

Longitudinal Change (3 Group) minimal frontal bizygomatic bitragal bigonial inner canthal outer canthal palpebral fissure upper facial depth mid facial depth lower facial depth nose length nasal bridge length philtrum length lower facial height total facial height ear length Age Group p 0. 4628 0. 0001 0. 0008 0. 0053 0. 0024 0. 0009 0. 0083 0. 0312 FAS vs. Deferred FAS vs. Control 0. 0001 0. 0003 0. 0014 0. 0046 0. 0037 0. 0139 0. 0036 0. 0009 0. 0166 0. 0424 Deferred vs. Control 0. 0072 0. 0542

difference = (visit 2 – visit 1)/visit 1 FAS deferred Control 0. 12 0. 1 0. 08 0. 06 0. 04 0. 02 0 upper facial mid facial lower facial depth

Conclusions from Comparisons n Many variables differ among the groups at each time point n n Few variables (bizygomatic width, facial depths) have significantly different rates of change across the two time points Relatively constant growth rate during this period

Replication n Growing sample of longitudinal data in San Diego 42 individuals in SD with 2 longitudinal images n Growing number are being seen in the next year for a longitudinal visit n

Plans for the Next Year n Collect images at 3 US sites n n Collect images in South Africa n n Small study in Minneapolis (increase # of US FAS) Both infants and adolescents Aim 1 – Delineate facial differences in alcohol exposure n Human only studies n n n Sample size to compare South Africa and African Americans More comparison with animal models Aim 2 - Longitudinal analyses n Currently cross camera; will start within camera

South Africa FAS Classification n FAS n n PFAS n n Short palpebral fissures, thin upper lip, flat/smooth philtrum, small OFC, growth retardation 2 of 3 facial + small OFC + growth retardation Heavy exposure n Mother drank ≥ 1 oz pure alcohol/day or 4 binges (≥ 4 drinks or 2 oz alcohol/occasion)