Early age of onset in 200 lymphoid cancer

Early age of onset in 200 lymphoid cancer families Samantha Jones, J Voong, J Graham, A Brooks-Wilson June 11 2019 1

Conflict of interest • The authors declare no conflict of interest. 2

Goals & objectives • Determine if familial lymphoid cancers are different from population cases, in terms of: • Sex distribution • Age of cancer onset • Determine if familial lymphoid cancers show anticipation (i. e. are diagnosed earlier in successive generations) 3

Background • Lymphoid cancer: • • • 6 th among all cancers worldwide 5 th in Canada Clinically and biologically heterogeneous group of neoplasms Familial lymphoid cancer is rare Elevated risk: • Relative with lymphoid cancer • Sex, ethnicity, viral/infectious agents 4

Lymphoid cancer family study • Ongoing family collection: 2006 – present • Recruitment: Physician referral, study website • Inclusion criterion: • Families with 2 or more lymphoid cancer cases • Lymphoid cancer cases and first degree blood relatives are interviewed and complete a standardized family history questionnaire • Collection of germline DNA, tumour DNA, and medical records • Tumour slides and tissue blocks are reviewed by an Oncology Pathologist to confirm the diagnosis 5

Familial Lymphoid Cancers • Types of lymphoid cancers: • • NHL, non-Hodgkin lymphoma HL, Hodgkin lymphoma CLL, chronic lymphocytic leukemia MM, multiple myeloma • Multigenerational • Heterogeneous 6

Characterization of families • 200 families; 527 lymphoid cancer cases • Most families were identified through a family member residing in British Columbia, Canada • 90% participants live in Canada • 7% in USA • 3% Other • Majority (73%) of study participants reported a subtype: • i. e. Diffuse large B-cell lymphoma is a subtype of NHL • Medical records and pathology slides available for 48% of cases • Confirmed the self-reported lymphoid cancer diagnosis in all 252 cases. 7

Distribution of lymphoma by sex • • NHL, p-value = 1 HL, p-value = 0. 0433 CLL, p-value = 0. 4340 MM, p-value = 0. 9380 Frequency (%) • Risk of lymphoma is higher in men than in women • Chi-square goodness-of-fit test • Familial HL cases were significantly different from, and less male than that of population cases 0. 0433 70 Population (male) Study (male) 60 50 40 30 20 10 0 NHL HL CLL MM Type of lymphoma All types 8

Race/ethnicity • Used SEER data because comparable data was unavailable for Canada • Race/ethnicity was classified according to SEER guidelines • Study participants: • 96% were non-Hispanic White • 3% were Asian • 1% don’t know or other 9

Age of onset varies by type • Mean age of diagnosis differs by type of lymphoid cancer • p-value <0. 0001 56 35 61 64 Mean age of diagnosis 10

Controlling for factors known to affect risk • Used SEER population data to generate cumulative incidence distributions (percentiles) • Percentiles control for: • • Sex Ethnicity/race Age of onset Subtype or type of lymphoma 11

Age of onset is earlier in families • Familial cases age of onset is earlier than that of population cases • Familial cases age of onset percentiles are earlier than population cases percentiles Number of cases and age of diagnosis by lymphoid cancer type Type of lymphoid cancer All cases NHL HL CLL MM Total Number of lymphoid cases 261 79 152 35 527 SEER 1 Study median age Years Earlier median age of diagnosis in Diagnosis of diagnosis (Percentile) 66 (50%) 39 (50%) 71 (50%) 69 (50%) 57 (28%) 31 (46%) 62 (26%) 64 (30%) 57 (29%) -9 -8 -9 -5 Sign median (age) p-values <0. 0001 0. 0065 <0. 0001 0. 1755 • Possible reason: Ascertainment bias • Systematic deviation from expected results due to recruitment methods 12

Ascertainment bias? 1. Remove cases diagnosed before 25 years old (reproductive age) Type of Number of lymphoid cancer cases All cases NHL 261 HL 79 CLL 152 MM 35 Total 527 Cases older than 25 NHL 230 HL 43 CLL 141 MM 33 Total 447 SEER 1 Study median age Years Earlier Sign median age of diagnosis in Diagnosis (age) of diagnosis (Percentile) p-values 66 (50%) 39 (50%) 71 (50%) 69 (50%) 57 (28%) 31 (46%) 62 (26%) 64 (30%) 57 (29%) -9 -8 -9 -5 <0. 0001 0. 0065 <0. 0001 0. 1755 66 (50%) 39 (50%) 71 (50%) 69 (50%) 57 (29%) 38 (63%) 62 (26%) 64 (30%) 58 (32%) -9 -1 -9 -5 <0. 0001 0. 4446 0. 0184 0. 6213 13

Ascertainment bias? 1. Remove cases diagnosed before 25 years old (reproductive age) 2. Remove probands Type of Number of lymphoid cancer cases All cases NHL 261 HL 79 CLL 152 MM 35 Total 527 Probands removed NHL 105 HL 45 CLL 61 MM 21 Total 232 SEER 1 Study median age Years Earlier Sign median age of diagnosis in Diagnosis (age) of diagnosis (Percentile) p-values 66 (50%) 39 (50%) 71 (50%) 69 (50%) 57 (28%) 31 (46%) 62 (26%) 64 (30%) 57 (29%) -9 -8 -9 -5 <0. 0001 0. 0065 <0. 0001 0. 1755 66 (50%) 39 (50%) 71 (50%) 69 (50%) 65 (42%) 36 (48%) 68 (44%) 68 (46%) 64 (45%) -1 -3 -3 -1 0. 0014 0. 0596 0. 0079 0. 4545 14

Ascertainment bias? 1. Remove cases diagnosed before 25 years old (reproductive age) 2. Remove probands 3. Remove short duration to develop disease Type of Number of lymphoid cancer cases All cases NHL 261 HL 79 CLL 152 MM 35 Total 527 Time to develop cancer NHL 96 HL 41 CLL 43 MM 16 Total 196 SEER 1 Study median age Years Earlier Sign median age of diagnosis in Diagnosis (age) of diagnosis (Percentile) p-values 66 (50%) 39 (50%) 71 (50%) 69 (50%) 57 (28%) 31 (46%) 62 (26%) 64 (30%) 57 (29%) -9 -8 -9 -5 <0. 0001 0. 0065 <0. 0001 0. 1755 66 (50%) 39 (50%) 71 (50%) 69 (50%) 60 (36%) 33 (42%) 61 (27%) 63 (30%) 58 (35%) -6 -6 -10 -6 0. 4884 0. 5515 0. 0273 0. 6636 15

Generations • Generation 1: First reported lymphoid cancer case, or suspected carrier status • HL 80 and NHL 75 are labeled generation 2 as we expect one or both parents to be carriers of a genetic factor • HL 14 is labeled generation 3 • CLL 35 is labeled generation 4 16

Anticipation • Anticipation is when disease severity increases and/or age of onset decreases in successive generations Age of diagnosis distribution by generation for NHL, CLL and MM n = 261 p<0. 0001 n = 79 p=0. 0015 n = 152 p=0. 0182 n = 35 p=0. 3649 17

Anticipation - Ascertainment bias? 1. Remove cases diagnosed before 25 years old (reproductive age) 2. Remove probands 3. Remove short duration to develop disease 18

Anticipation - Ascertainment bias? NHL 1. Cases diagnosed before 25 years old (reproductive age) 2. Probands 3. Short duration to develop disease Removed All NHL cases 1. Cases dx ≥ 25 2. Probands 3. Short duration n 261 130 105 96 median (percentile) 57 (28%) 57 (29%) 65 (42%) 60 (36%) p-value <0. 0001 0. 0013 19

Anticipation - Ascertainment bias? HL 1. Cases diagnosed before 25 years old (reproductive age) 2. Probands 3. Short duration to develop disease Removed All HL cases 1. Cases dx ≥ 25 2. Probands 3. Short duration n 79 43 45 41 median (percentile) 31 (46%) 38 (63%) 36 (53%) 33 (42%) p-value 0. 0015 0. 2840 0. 0573 0. 0941 20

Anticipation - Ascertainment bias? CLL 1. Cases diagnosed before 25 years old (reproductive age) 2. Probands 3. Short duration to develop disease Removed All CLL cases 1. Cases dx ≥ 25 2. Probands 3. Short duration n 152 141 61 43 median (percentile) 62 (26%) 68 (44%) 61 (27%) p-value <0. 0001 0. 0184 0. 3221 0. 4400 21

Anticipation - Ascertainment bias? MM 1. Cases diagnosed before 25 years old (reproductive age) 2. Probands 3. Short duration to develop disease Removed All MM cases 1. Cases dx ≥ 25 2. Probands 3. Short duration n 35 33 21 16 median (percentile) 64 (30%) 68 (46%) 63 (30%) p-value 0. 3649 0. 3741 0. 3317 0. 7068 22

Anticipation - Ascertainment bias? All types 1. Cases diagnosed before 25 years old (reproductive age) 2. Probands 3. Short duration to develop disease Removed All cases 1. Cases dx ≥ 25 2. Probands 3. Short duration n 527 447 232 196 median (percentile) 57 (29%) 58 (32%) 64 (45%) 60 (35%) p-value <0. 0001 0. 0023 23

Conclusions • Mean age of onset differs by type, as expected • Median familial NHL, CLL (but not MM) ages of onset are substantially earlier than population data • Observe anticipation in familial cases of NHL, HL and CLL (but not MM) • For NHL and lymphoid cancers collectively, anticipation does not appear to be due to ascertainment bias • Larger sample size for HL, CLL and MM could help determine more clearly if ascertainment bias is present • Future direction: Inclusion of additional lymphoid cancer families with collaborators 24

A word of thanks Simon Fraser University: Jinko Graham Jackson Voong Participants of the Lymphoid Cancer Families Study BC Cancer: Joe Connors Graham Slack Amy Williams Stephen Leach My Ph. D Supervisory Committee: Angela Brooks-Wilson Carolyn Brown Inanc Birol Jan Friedman 25 25

Characterization of families • 200 families; 527 lymphoid cancer cases NHL HL CLL MM All cases # of cases 261 (50%) 79 (15%) 152 (29%) 35 (7%) 527 Sex (male) 145 (56%) 34 (43%) 84 (55%) 18 (51%) 281 (53%) Dx. confirmed 121 (46%) 36 (46%) 81 (53%) 14 (40%) 252 (48%) Ethnicity White Asian Other 247 (95%) 13 (5%) 1 (1%) 76 (96%) 3 (4%) - 150 (99%) 2 (1%) 34 (97%) 1 (3%) - 507 (96%) 17 (3%) 3 (1%) Location Canada USA Europe Other 229 (88%) 20 (8%) 10 (4%) 2 (1%) 67 (85%) 8 (10%) 3 (4%) 1 (1%) 144 (95%) 6 (4%) 2 (1%) - 29 (83%) 3 (9%) 2 (6%) 1 (3%) 469 (89%) 37 (7%) 17 (3%) 4 (1%) 26