Comprehensive Somatic Copy Number Analysis Using Aqueous Humor Liquid Biopsy for Retinoblastoma

Abstract

Aqueous humor (AH) liquid biopsy has been established as a surrogate tumor biopsy for retinoblastoma (RB). Previous AH studies have focused on highly recurrent RB somatic copy number alterations (SCNAs) including gain of 1q, 2p, 6p, and loss of 13q and 16q. In this retrospective study, we provide a comprehensive, whole-genome analysis of RB SCNAs and evaluate associated clinical features for 68 eyes of 64 RB patients from whom AH was obtained between December 2014 and October 2020. Shallow whole-genome sequencing of AH cell-free DNA was performed to assess for SCNAs. The prevalence of specific non-highly recurrent SCNAs, such as 20q gain and 8p loss, differed between primarily and secondarily enucleated eyes. Increases in chromosomal instability predict more advanced seeding morphology (p = 0.015); later age of diagnosis (p < 0.0001); greater odds of an endophytic tumor growth pattern (without retinal detachment; p = 0.047); tumor heights >10 mm (p = 0.09); and containing 6p gain, a biomarker of poor ocular prognosis (p = 0.004). The AH liquid biopsy platform is a high-yield method of whole-genome RB SCNA analysis, and SCNAs are associated with numerous clinical findings in RB eyes. Prospective analyses are encouraged to further elucidate the clinical relevance of specific SCNAs in RB.

Keywords: SCNA; aqueous humor; cell-free DNA; circulating tumor DNA; liquid biopsy; precision oncology; prognostic biomarker; retinoblastoma.

Figure 1

Composite SCNA plot for all eyes included in our study. Gains are represented by the red line and losses are represented by the blue line. Besides previously determined highly recurrent RB SCNAs of 1q, 2p, and 6p gain as well as 16q loss, non-highly recurrent SCNAs were also seen. The most common were 12p loss, 16p loss, 5p gain, 17q gain, 18q gain, 17p loss, 20q gain, and 22p gain.

Figure 2

Box and whisker plot showing (A) genomic instability based on seeding classification and (B) integer numbers of SCNAs based on seeding classification. x indicates the median value.

Figure 3

Box and whisker plot showing genomic instability based on laterality, heritability (+ RB1 indicates a germline mutation), growth pattern, and height. x indicates the median value.

Figure 4

The prevalence of specific non-highly recurrent SCNAs differed between primarily and secondarily enucleated eyes. Note, not all secondarily enucleated eyes had AH sampled prior to enucleation to be used for comparison of new SCNAs.

Figure 5

Four eyes demonstrated new SCNAs at secondary enucleation that were not present during conservative management. New SCNAs were as follows: Case 11 (8q loss, 16q loss, 20q gain), Case 15 (1q gain, 6p gain, 17p loss, 19p gain), Case 17 (1p loss, 1q gain, 4p loss, 4q loss, 5p gain, 5q gain, 8p loss, 8q loss, 12p loss, 12q loss, 15q loss, 16p loss, 16q loss, 18p gain, 18q gain, 19p gain, 19q gain, 20p gain, 20q gain, 21q gain), and Case 33 (2p gain, 19q loss). Other focal changes seen are at the centromeres and not true SCNAs. The first samples taken from cases 11, 13, and 17 were during IVM injection, whereas the first sample taken from case 33 was at diagnosis. Treatment courses for these cases are as follows. Case 11 underwent systemic chemotherapy treatment (six cycles of carboplatin, etoposide, and vincristine (CEV)), followed by four IVM injections over three months for recurrent dust seeding; the eye was enucleated 14 months after diagnosis due to primary tumor recurrence. Case 15 began treatment with three monthly intra-arterial melphalan injections, followed by three weekly IVM injections for dust seeding; the eye was enucleated 22 months after diagnosis due to primary tumor recurrence. Case 17 underwent systemic chemotherapy treatment (six cycles CEV), followed by three IVM injections over one month for recurrent sphere seeding; the eye was enucleated two months later due to primary tumor recurrence. Finally, Case 33 was treated with two cycles of CEV as a bridge and three cycles of intra-arterial melphalan, followed by four IVM injections over six weeks to treat persistent dust seeding; the eye was enucleated six months after diagnosis due to apical tumor recurrence with persistently active seeding despite ongoing therapy.