Cell-free RB1 DNA not detected in the blood of pseudoretinoblastoma patients

Introduction

Cell-free DNA (CfDNA) is now a clinical tool used to aid the diagnosis, prognosis and management of diverse cancers.1 We have demonstrated that RB1 cfDNA is commonly detected in the plasma of retinoblastoma patients before treatment and in the one patient studied the level increased after a month without treatment.2 Following enucleation or intraarterial chemotherapy, it becomes undetectable.2 3

The diagnosis of retinoblastoma is usually made with the indirect ophthalmoscope and sometimes aided with ultrasound and MRI.4 Biopsy is not done because of the concern for spreading the cancer. A diverse group of intraocular lesions that resemble retinoblastoma (but are not cancer) simulate retinoblastoma and accurate diagnosis at times is very difficult and errors continue to be made resulting in inappropriate surgery and even removal of eyes in children who do not have cancer.5–7 In addition, RB1 abnormalities have been reported in a variety of other conditions including rhabdomyosarcoma.8

Tumour-specific plasma cfDNA is now being used in adult and paediatric solid cancers for diagnosis, differential diagnosis, prognosis, monitoring of response and detection of tumour mutations. We have shown that RB1 fragments (ctDNA) are commonly (60%–80%) present in the blood (plasma) of children with newly diagnosed retinoblastoma.2 9 10 The purpose of this study was to see if RB1 fragments are detectible in the plasma of patients with pseudoretinoblastoma.

Methods

The cfDNA was analysed with hybridisation capture and next-generation sequencing in blood (plasma) using MSKCC’s analysis of circulating cfDNA to evaluate somatic status11 in 40 patients with diverse pseudoretinoblastoma lesions. This technique interrogates 129 established cancer mutations and because the buffy coat is simultaneously analysed germline defects and clonal haematopoiesis are filtered out in the results. Children with pseudoretinoblastoma like lesions that had been referred to our centre for a possible diagnosis of retinoblastoma were included.

Results

Table 1 depicts/shows that for 40 patients with pseudoretinoblastoma, plasma RB1 cfDNA was not detectable in any patient.

Discussion

The diagnosis of retinoblastoma is usually made with the indirect ophthalmoscope and often ultrasound and MRI are performed. There are more than 30 conditions that have been described that mimic the appearance of retinoblastoma and they are referred to as ‘pseudoretinoblastomas.’ The most common ‘pseudoretinoblastomas’ are Coats’ disease, persistent fetal vasculature syndrome, retinal detachment, ocular toxocariasis and other infectious, genetic and even traumatic lesions (table 1).1 Despite the assistance of ultrasound and MRI, the differential diagnosis of retinoblastoma remains a challenge and errors are still made with children’s eyes receiving inappropriate surgery including vitrectomy, retinal detachment repair and enucleation.5–7 In addition, some eyes with retinoblastoma are not enucleated because the clinician misdiagnosed them as benign conditions. The true, modern incidence of these errors is not known because few centres are eager to publish on their errors in diagnosis, but recent publications suggest an inappropriate enucleation rate in some centres of 11%–24%.5–7

MSKCC developed a next generation sequencing assay for 129 cancer-related genetic abnormalities and one of these genes is RB1. All exons of RB1 are interrogated.

In this study, we assayed for cfDNA of RB1 and in all cases no RB1 fragments were detectable. These assays were done simultaneously with true retinoblastoma samples (which did detect RB1 abnormalities) and occasionally other, non-RB1 abnormalities giving us additional confidence that the assay was accurate. This suggests that finding RB1 cfDNA in the blood of a child with suspected retinoblastoma is not a pseudoretinoblastoma. Whether this test can be used as an aid in the differential diagnosis of retinoblastoma will depend on additional analyses of sensitivity and specificity in a larger cohort of both retinoblastoma and pseudoretinoblastoma eyes, but neither this study nor any prior published studies have demonstrated RB1 abnormalities in pseudoretinoblastomas.