Papillary Adenocarcinoma of the Iris Transmitted by Corneal Transplantation

To our knowledge, there have been no reports of transmission of a systemic malignancy from a donor to a recipient by corneal transplantation. Known cases of transmission of disease by transplanted corneal tissue have involved infectious agents^1-14 and 1 case of intraocular tumor.¹⁵ Corneal tissue from donors who have died of malignancy (with the exception of leukemia, lymphoma, and retinoblastoma) has been considered safe for transplantation.^15-19 In contrast, there have been many reports of transmission of systemic malignancy from donor to recipient by solid organ transplantation. The donor origin of one case of tumor transmission by solid organ transplantation was proven conclusively by DNA typing.²⁰ We report the first known instance, to our knowledge, of transmission of a systemic malignancy by corneal transplantation, proven by DNA analyses.

A 22-year-old man sought treatment in June 1992 because of a 2-week history of slight irritation of his right eye and an area of discoloration of his right iris. He had undergone a right corneal graft for keratoconus in November 1990 (19 months previously). The left eye was densely amblyopic and had not been grafted. At a review 2 months prior to the June 1992 visit, the right eye was quiet, with no evidence of a mass, and a corrected visual acuity of 20/20.

Two months later, the visual acuity in the right eye was 20/40 with correction, improving with pinhole to 20/20. There was a vascularized, nonpigmented mass approximately 4.0 × 2.5 mm, arising from the inferotemporal iris between the 6- and 8-o'clock positions and extending into the angle (Figure 1A). The eye was inflamed, with dilated episcleral vessels adjacent to the mass, keratic precipitates, anterior chamber cells, and an inferotemporal posterior synechia. Indented funduscopy revealed a localized retinal dialysis but no evidence of ciliary body involvement. There was no evidence of posterior segment involvement on ultrasound and computed tomography. Results of a full clinical examination were unremarkable, and an intensive screen for systemic malignancy was negative. The provisional diagnosis was a granulomatous reaction to a foreign body, although malignancy could not be excluded. During the next week, the mass grew rapidly and it was decided in consultation with the patient and his family to proceed to excisional biopsy.

A partial-thickness scleral incision was made 2 mm from the limbus and a flap was raised, extending into clear cornea. A block of inner sclera including the scleral spur and the trabecular meshwork was excised. The iris was then prolapsed into the wound and a sector including the mass, the adjacent iris root, the face of the ciliary body, and the anterior ciliary processes was removed. The incision was closed and cryotherapy was applied to the posterior wound margin and the area of retinal dialysis. The immediate postoperative course was uneventful apart from a transient hyphema. Three weeks after the operation, the right eye was quiet and the visual acuity was 20/30 with pinhole.

Histologic examination of the iris mass revealed a poorly differentiated adenocarcinoma with associated inflammation. In view of the narrow margins around the mass and the probability of a persistent tumor in the adjacent angle, the patient's right eye received 10 000 rad (100 Gy) of ¹²⁵iodine plaque radiotherapy to a depth of 3 mm at the limbus from 4- to 10-o'clock position. At the time of removal of the plaque, a scleral buckle was placed at the site of the retinal dialysis.

At subsequent follow-up, there was a transient fall in visual acuity to 20/60 due to decreased clarity of the corneal graft (Figure 1B). The graft gradually cleared, and at the most recent follow-up nearly 10 years after the initial examination, the best-corrected visual acuity had improved to 20/30. There had been no evidence of primary adenocarcinoma in the recipient at any stage during this follow-up and no evidence of metastasis.

The corneal donor's medical records were retrieved and reviewed. The donor had impaired vision in both eyes and it was established that an ophthalmic examination had been performed in the weeks prior to death. The results of that examination had revealed bilateral choroidal masses consistent with choroidal metastases. The donor died in November 1990 of disseminated, poorly differentiated adenocarcinoma. The primary tumor was thought to originate from the bowel. An autopsy was not carried out but a percutaneous biopsy of a lung lesion had been performed 6 months prior to the donor's death. The histologic examination of the specimen had showed adenocarcinoma. A small sample of this specimen was still available and, together with the lesion from the recipient's iris and blood sample, was submitted for genetic analysis. Results of molecular analysis suggested that the recipient's iris tumor had arisen from the corneal donor.

The recipient of the donor's other cornea was traced and examined for evidence of tumor. The grafted eye was quiet and clear, with no evidence of tumor transmission. This recipient remained well in this and subsequent follow-up. No other tissues from the corneal donor were used for transplantation.

The biopsy tissue was submitted for routine pathologic examination. For light microscopy, the formalin-fixed tissue was embedded in paraffin and sections were stained with hematoxylin-eosin and a variety of specific commercial immunohistochemical stains.

The DNA was extracted from the tissue samples by incubation at 55°C in the presence of sodium dodecyl sulfate and proteinase K.²¹ Two extractions in phenol-chloroform were followed by a precipitation of the purified DNA in absolute alcohol at −70°C. Patient and donor tissue samples were extracted from paraffin blocks (fresh or fixed unprocessed tissue not required). DNA was amplified and typed at the DQα locus using the commercial Roche molecular systems typing kit (F. Hoffman La Roche Ltd, Basel, Switzerland). The DQ locus on chromosome 6 contains the genes that encode for HLA class II (HLA-D).²² There are 6 common DQ alleles detected by the kit (DQ 1.1, 1.2, 1.3, 2, 3, 4) that determine 21 possible genotypes. Appropriate negative (water) and positive (known 1.1, 4) controls were used. The reaction mix was amplified using 32 cycles of 94°C for 1 minute, 60°C for 30 seconds, 72°C for 30 seconds, and a final elongation of 72°C for 8 minutes. Amplified DNA was typed using the probing strips with 9 probes on their surfaces²³ and a hybridization temperature of 55°C for 20 minutes.

The specimen consisted of a segment of iris and ciliary body structures and the mass, which measured approximately 5 × 3 × 3 mm (Figure 2). Arising from the anterior surface of the iris and extending into the loose iris stroma was a malignant neoplasm with well-developed papillary architecture. The tumor comprised fibrovascular cores surrounded by focally stratified large epithelial cells with lightly eosinophilic cytoplasm, large irregular ovoid nuclei, and 1 or 2 prominent nucleoli. The cells were nonpigmented and occasional cells were vacuolated. Deeper parts of the tumor contained small solid nests of cells and occasional multinucleated cells. There was a high mitotic rate but no necrosis. The scleral spur was sectioned separately and showed occasional tumor cells adherent to the endothelium and trabecular meshwork but no scleral invasion. The tumor cells stained for cytokeratin but were negative for S100 protein. The tumor was diagnosed as a poorly differentiated adenocarcinoma. Further histologic stains for placenta-like alkaline phosphatase and α-fetoprotein (for testicular carcinoma) and thyroid globulin (for thyroid carcinoma) were negative.

Genetic analysis of the recipient revealed alleles 1.1 and 1.3. The donor analysis revealed alleles 1.2 and 3. The alleles detected in the tumor sample contained both alleles from the recipient (1.1, 1.3) and allele 3, also found in the biopsy sample from the donor (Table 1 and Figure 3). Because of the nature of this testing kit, the 1.2 allele can be hidden in this configuration; thus, its presence cannot be excluded or confirmed in the tumor specimen. To confirm the presence of the 1.2 allele in the iris tumor, DNA sequencing would be required. This was precluded by the small initial sample size.

The results of the molecular analyses almost conclusively demonstrate that the iris tumor, a poorly differentiated adenocarcinoma, arose from the corneal donor. The presence of the 3 allele in the tumor specimen can only be explained by the presence of donor cells in the tumor. The 1.2 allele was not detected in the tumor because of the typing kit methods. The tissue-typing kit is designed to detect 6 common DQ alleles. In normal circumstances, only 2 alleles would be present, and the kit relies partly on the exclusion of possible alleles for full typing (Table 1 and Figure 3). Therefore, because there could have been as many as 4 alleles, the kit was not able to confirm or exclude the presence in the iris tumor of the 1.2 allele from the donor.

The only alternative explanation for the presence of the 3 allele in the iris tumor sample is that there was a mutation in the recipient's iris tissue of either the 1.1 or 1.3 allele to a configuration that could be read by the 3 probe. Such a mutation would require 5 specific and independent replacement mutations, with the statistical probability that this could happen by chance in less than 1 in 10 million.

Transmission of systemic malignancy by solid organ transplantation has been recognized for many years.^24-39 The Cincinnati Transplant Tumor Registry reported 142 cadaver organ donations from patients with unrecognized malignancies, leading to 64 cases of recipient disease (45%).³⁸ Twenty-six (72%) of 36 recipients with distant metastases died of their tumors. Stringent criteria and procedures have been introduced in an attempt to reduce the incidence of this often-fatal complication but occasional cases are still reported.^36,39 The tumors most likely to evade detection are small and clinically silent but capable of early metastasis.^29,38

The use of closely matched donors and the systemic immunosuppression required for the survival of solid organ transplantations is thought to increase the viability of an inadvertently transmitted malignancy.^24,25,29,36 There have been several cases of successful treatment of a transplanted tumor by removing the transplanted organ, withdrawing immunosuppression, and in some cases, adding chemotherapy or radiation.^{24,25,27,30,34,39} In a previous report, a 57-year-old woman received a closely matched (6 of 6 HLA antigen matched) renal transplant, but 3 months later began to develop complications from what proved to be a malignant melanoma originating from the kidney donor. She died despite treatment, and an autopsy revealed a widely disseminated tumor. The large amount of tissue available made identification of the HLA-DQ alleles and several other polymorphisms feasible. These showed that despite the close antigen match, the donor and tumor differed from the recipient in 1 HLA-DQ allele. Three of 4 polymorphisms also showed that the donor and tumor were identical but differed from the recipient.²⁰

Our case represents the first documented case of transmission of a systemic malignancy by corneal transplantation. The clarity, avascularity, and small mass of the cornea make the accidental transmission of a viable number of tumor cells very unlikely, and with the exception of 1 case of transmission of retinoblastoma by corneal transplantation reported more than 60 years ago,¹⁵ this has never been known to occur. Similarly, there has been no evidence of tumor transmission by transplantation of corneas from donor eyes with primary choroidal melanomas.⁴⁰ In addition, corneal graft recipients very rarely receive systemic immunosuppression. For these reasons, the exclusion criteria for corneal donors have not been as strict as those for other organ donors.^2,17,18

However, the avascularity of the cornea and anterior chamber and the postoperative use of topical steroids may favor tumor transmission. This risk was raised by Zakov et al,⁴¹ who described a case in which micrometastases were found close to the excision margin of the corneoscleral button in a donor eye. However, there was no evidence of transmission of malignancy in this case or in a retrospective analysis of 403 cases of corneal transplantation at the Massachusetts Eye and Ear Infirmary (Boston) between 1965 and 1968.¹⁶ Since these reports, it has been considered safe to use such donors for corneal transplantation,^17,18 with further support from a recent retrospective review of 143 patients.¹⁹ However, direct evidence of the malignant potential of micrometastatic tumor cells has been shown.⁴² Furthermore, anterior chamber–associated immune deviation at least theoretically might explain how situations such as our patient's arise. Experimentally it has been shown in mice that immune privilege is extended to foreign tumor cells in the anterior chamber. These foreign tumor cells survive and progressively grow in the immune deviant anterior chamber environment.⁴³

In our case, the finding of bilateral choroidal masses in the donor during a clinical examination makes it highly likely that there was metastatic adenocarcinoma in both eyes at the time of death. Presumably there were small numbers of malignant cells adherent to the corneal endothelium at the time of harvesting, and some or all of these cells proliferated in the recipient anterior segment. To date there has been no evidence of extraocular spread.

It is interesting that the tumor did not become apparent for 19 months after the date of transplantation, despite the apparent rapid growth at presentation and the poorly differentiated histologic characteristics. It is well known that even some very aggressive tumors, such as poorly differentiated breast adenocarcinomas, may not recur for up to 5 years after excision of the primary tumor.⁴⁴

Another factor that may have played a part in this case is that the original implanted tumor cell(s) may not have had a blood supply on the avascular graft endothelium. Their growth may have been slow until a cell reached a vascular supply on the iris or trabecular meshwork. The recipient's immune system may have partially controlled the tumor while it was small, losing control as it grew bigger.

Evidence from patients with a solid organ tumor who have developed transplanted malignancies shows that it is possible for a competent immune system to overcome such a malignancy if the tumor load is not too great.^{24,25,27,30,34,39} Such immune modulation of tumor growth is thought to be mediated by natural killer–mediated mechanisms.⁴⁵ The recommended management is to remove as much of the tumor as possible, withdraw immunosuppression, and consider the use of adjunctive radiation or chemotherapy.^{24,25,27,29,30,34,38} The graft recipient in this report had a competent immune system and a poorly matched donor, unlike most solid organ transplant recipients. There was evidence of an immune response to the tumor clinically, prior to excisional biopsy, and on histologic examination. After excision cryotherapy⁴⁶ and local radiation, the residual tumor load was almost certainly negligible if not nil. After nearly 10 years of careful follow-up, the recipient has remained free of problems. Although we continue to be vigilant, we are increasingly confident that the tumor has been "cured" and that the patient's immune response has eliminated any remaining tumor cells.

Eye bank procedures currently exclude all donors with proven^1-10,15 or likely potential to transmit disease to recipients.^{12,17,18,47-49} Although this case demonstrates a new risk of transmission of donor disease to a recipient, it should be interpreted with caution. As many as 40% of donors in most eye banks have died of disseminated malignant tumors and therefore have potential micrometastases to the eye.⁴¹ Despite these large numbers of donors, no case of transmission has previously been noted.^16,19

At this stage, on the evidence of 1 case, the risk of such transmission must be very low. Even if transmission occurs, the factors of probable poor tissue match, lack of immunosuppression, and small tumor load make the prognosis for patient survival excellent and for retention of the eye good. It would clearly be prudent, however, to exclude potential donors with known ocular metastases. All recipients need careful long-term follow-up and a high degree of suspicion for malignancy in unusual iris presentations.

The authors have no financial interest in any of the instruments or techniques used in this article.

We acknowledge the assistance of Jerry A. Shields, MD, and Ralph Eagle Jr, MD (Wills Hospital, Philadelphia, Pa) and Derek Sherwood, FRCOphth (Nelson Hospital, New Zealand) in the treatment of this patient, and the resources of the New Zealand National Eye Bank.

Corresponding author and reprints: Phillip J. Polkinghorne, FRCOphth, Department of Ophthalmology, Auckland Hospital, Private Bag 92024, Auckland, New Zealand (e-mail: philip@pjpolk.co.nz).