Question 1A How to Make the Diagnosis of Primary Uveal Tract

Question 1a How to make the diagnosis of primary uveal tract melanoma? Data base: Second sift from 722 papers related to primary treatment of uveal melanoma Inclusion Criteria: Human studies, case series >5, phase II trails, all RCTS’s, Cohort studies, Case control studies Results: 25 of 722 relevant to answer Question 1a (refined to 12)

Collaborative Ocular Melanoma Study Group. Accuracy of diagnosis of choroidal melanomas in the Collaborative Ocular Melanoma study. Arch Ophthalmol 1990;108:1268–73.

RCT early paper in recruitment phase

In 3-year period from 1989, 413 patients diagnosed with UM using ophthalmoscopy, photography, conventional ocular ultrasonography. Histology available on all patients. Misdiagnosis rate 0.48%

 Ophthalmoscopy (0)  Photography (0)  Ultrasound techniques (6) includes comparison with PET/CT and AS-OCT  Intraocular biopsy techniques (4)  Comparative study (1)

1 Ocular Ultrasound and UBM

1. Characteristic Ultrasonographic Findings of Choroidal Tumors. Wang et al Journal of Medical Ultrasound 2003

Study type Retrospective case notes review Aims To describe the ultrasound features of different choroidal tumours Inclusion criteria Not detailed Exclusion criteria Not detailed Method B scan 10 MHz probe and A scan though closed eye lid Number of patients 27 Outcomes 10 choroidal melanoma described All classic collar stud shape 50% medium reflectivity and 40% low to medium reflectivity

Conclusions Well described B scan and A scan ultrasound characteristics can help to refine the type of intraocular tumour Comment Useful noninvasive investigation Very small collection of for choroidal tumours cases from Taiwan Unverisity hospital over 4 years!

Purpose: To report the characteristic ultrasonographic findings of choroidal tumors diagnosed at a university hospital between 1999 and 2002.Materials and Methods: The charts of patients with choroidal tumors diagnosed between January 1999 and December 2002 were reviewed retrospectively. Age, gender and symptoms were recorded. The characteristic ultrasonographic findings of choroidal tumors, including shape, internal reflectivity, location and associated retinal detachment, were analyzed.Results: A total of 27 cases of choroidal tumors were reviewed. Choroidal hemangioma was found in 15 cases (56%), choroidal melanoma in 10 (37%), and metastatic choroidal tumor in two (7%). Fourteen patients had a dome-shaped tumor mass (14 choroidal hemangiomas), 11 had a collar-button tumor (10 choroidal melanomas and 1 choroidal hemangioma), and two had an irregular/bumpy tumor (metastatic choroidal tumors). Retinal detachment was noted in 14 cases. Thirteen choroidal hemangiomas (86.7%) were located adjacent to the optic nerve, while eight choroidal melanomas (80%) were not located at the posterior pole. High internal reflectivity was noted in 13 choroidal hemangiomas (86.7%). Five cases of choroidal melanoma (50%) had medium to high internal reflectivity, and four cases (40%) showed low to medium internal reflectivity.Conclusions: Ultrasonography is a noninvasive examination for choroidal tumors. However, not all choroidal tumors had the typical pictures described by previous studies. From our study, the shape, location, reflectivity, and associated retinal detachment might be helpful indicators for differential diagnosis of choroidal tumors.

2 2. The Use of Ultrasound Biomicroscopy in the Evaluation of Anterior Segment Tumors and Simulating Conditions Gündüz et al Department of Ophthalmology, Ankara University Faculty of Medicine, Ankara, Turkey Ophthalmologica 2007;221:305–312. Study type Case series, partly comparative Aims Report UBM using 50MHz probe for various anterior segment diagnoses some confirmed on histopathology Inclusion criteria Any anterior segment lesion Exclusion criteria Not stated Method Prospective analysis of data from Aug 2002-2006 Comparison with histology where available Number of patients 35 patients 7 Ciliary Body MM 4 Iris MM 2 Ring Melanoma Outcomes Ciliary body MM had low to medium reflectivity Of the 6 CB tumours reported there was no correlation between UBM finding and histological subtype Conclusions Good for detecting small CB MM and to differenciate iris naevus from iris melanoma and identify pigment epithelial cysts Comment Useful sign loss of the acute angle shape in ring melanoma

Abstract

Purpose: To report the ultrasound biomicroscopy (UBM) findings of anterior segment tumors and simulating conditions. Methods: Thirty-five patients underwent UBM. Of those, 16 had histopathologically or cytopathologically diagnosed tumors, and 19 had clinically diagnosed lesions. Re- sults: The study material comprised 13 iris pigment epithelial (IPE) cysts, 7 ciliary body melanomas, 4 iris melanomas, 4 iris nevi, 3 intraocular invasions of conjunctival squamous cell carcinoma, 2 ring melanomas of the anterior chamber angle, 1 medulloepithelioma and 1 pars plana cyst. On UBM, all IPE cysts presented as cystic lesions with a thin cyst wall and no solid components. All ciliary body melanomas showed low to medium reflectivity, with cavitation in one case and extraocular extension in another. Iris melanomas presented as anterior (stromal) iris lesions with medium to high internal reflectivity. There was irregularity and convex bowing of the posterior iris plane in iris melanomas, a fea- ture not seen in iris nevi. Intraocular invasion of conjunctival squamous cell carcinoma was evidenced as areas of medium to high reflectivity in the ciliary body and iris, loss of the acute angle shape and highly reflective spots in the anterior chamber. Conclusions: UBM was particularly useful in the diagnosis of IPE cysts, in the visualization of small ciliary body melanomas, in the differentiation of iris melanomas from iris

3 nevi and in the demonstration of intraocular invasion from conjunctival squamous cell carcinoma. Ultrasound biomicroscopy (UBM) provides high-resolution in vivo imaging of the anterior segment in a noninvasive fashion. In addition to imaging the tissues easily seen by clinical methods, it images structures otherwise hidden from clinical visualization, including the ciliary body and zonules, allowing assessment of their morphol- ogy [1–3]. The purpose of this study was to evaluate the UBM features of anterior segment tumors and simulating conditions.

4 3. Assessment of Anterior Segment Tumors with Ultrasound Biomicroscopy versus Anterior Segment Optical Coherence Tomography in 200 Cases. Carlos Bianciotto et al. Ophthalmology 2011;118:1297–1302

Study type Retrospective non-interventional case series Aims Compare UBM with OCT imaging to determine which is best Inclusion criteria All anterior segment tumours Not specific to melanoma Exclusion criteria Method Number of patients 200 47 (24%) had melanomas Outcomes UBM was more favorable for resolution of the posterior margin of the lesion and for structures from the pigment epithelium posteriorly, whereas AS-OCT was more favorable for anterior margin and ocular structures anterior to the IPE

Image quality was considered good with UBM in 80%versus 68% with AS-OCT

Overall, UBM was more favorable for complete tumor resolution in IPE cyst and iris melanoma

Conclusions AS-OCT suffers from optically-related image shadowing with large, pigmented, IPE and ciliary body lesions.

Advantage of AS OCT it is quick and easy to perform and not uncomfortable for the patient ( no waterbath is required) It is useful iris MM anterior to the PE for and conj tumours

Comment UBM is better than AS OCT for anterior segment uveal melanoma

Abstract

5 Purpose: To compare ultrasound biomicroscopy (UBM) versus anterior segment optical coherence tomography (AS-OCT) for imaging of tumors of the anterior segment of the eye.Design: Retrospective, noninterventional case series. Participants: We included 200 patients. Methods: Review of medical records of patients who underwent both UBM and AS- OCT for evaluation of anterior segment tumors. Main Outcome Measures: Comparison of tumor surface and internal visualization.

Results: There were 200 eyes with anterior segment tumors involving the iris stroma in 96 (48%), ciliary body in 14 (7%), combined iris and ciliary body in 32 (16%), iris pigment epithelium (IPE) in 44 (22%), conjunctiva in 6 (3%), sclera in 4 (2%), and others in 6 (1% each). The diagnoses included nevus in 75 eyes (38%), melanoma in 47 (24%), cyst in 48 (24%), epithelioma (adenoma) in 5 (3%), metastasis, melanocytosis and melanocytoma in 4 eyes each (2%), and others (1% each). Image analysis (UBM vs AS-OCT) revealed adequate visualization of all tumor margins (189 [95%] vs 80 [40%]), posterior tumor shadowing (9 [5%] vs 144 [72%]), and high overall image quality (159 [80%] vs 136 [68%]). Comparison for better image resolution (UBM vs AS-OCT) disclosed UBM provided better overall tumor visualization (138 [69%] vs 62 [31%]) and better resolution of the posterior margin (147 [74%] vs 53 [27%]), whereas AS-OCT provided better resolution of the anterior margin (40 [20%] vs 160 [80%]) as well as better overall resolution of anterior segment anatomy (41 [21%] vs 159 [80%]). Better resolution was found with UBM for pigmented tumors (n � 162; 107 [66%] vs 55 [34%]) as well as for nonpigmented tumors (n � 38; 23 [61%] vs 15 [39%]). Regarding location, iris tumor resolution was similar with each technique (49 [52%] vs 45 [48%]).

Conclusions: For anterior segment tumors, UBM offers better visualization of the posterior margin and provides overall better images for entire tumor configuration compared with AS-OCT.

7 4. Anterior segment imaging for iris melanocytic tumors. Razzaq et al European Journal of Ophthalmology, 2011;21(5): 608-614

Study type Retrospective comparative cohort Aims AS-OCT verses UBM for Inclusion criteria Melanocytic iris lesions Exclusion criteria Method Image quality was compared Not all OCT machines used with UBM as gold standard are widely available and between 2006 and 2009 therefore this limits usefulness of this paper Number of patients 61 ASOCT was compared with UBM in 42 patients Outcomes AS OCT compared well with UBM in 86% of cases

But AS OCT only detected CB extension of iris melanoma in 1 of 3 cases.

Conclusions UBM is superior for detecting CB extension Comment Precise tumour measurement could be made with AS OCT to allow for radiotherapy planning Abstract

Purpose: To determine the role of anterior segment Optical Coherence Tomography (AS-OCT) and other anterior segment imaging techniques (Pentacam and Slit lamp- OCT) for analysis of iris melanocytic tumors and to compare results with clinical features and ultrasound biomicroscopy (UBM). Methods: Between 2006 and 2009, sixty- one patients with melanocytic iris tumors were examined using a variety of anterior segment imaging techniques, i.e. Pentacam, Slit lamp-OCT (SL-OCT), Anterior segment OCT (AS-OCT) and UBM (50MHz). Pentacam was performed in 17 patients, SL-OCT in 12, AS-OCT in 46 and UBM in 49 patients.Results: The Pentacam images identified the tumor in three of 17 patients (18%), SL-OCT in eight of 12 (67%) and AS-OCT in 44 of 46 patients (96%). AS-OCT results were compared with UBM in 42 patients: in 86% of these, the results were comparable, although AS-OCT demonstrated a ciliary body extension of iris tumors in only one out of three cases analyzed.Conclusion: Iris melanocytic tumors were located by AS-OCT in 96% of cases and results were comparable to UBM imaging, while both SL-OCT and Pentacam were less reliable for detecting and measuring anterior segment lesions. AS-OCT gives precise anterior eye segment diameters, which are useful for calculating brachytherapy dosage using a module developed at the LUMC. Therefore, AS-OCT is a reliable, convenient and non-contact method for demonstrating and measuring pure iris tumors, but UBM is superior in detecting a ciliary body extension of these tumors.

5. Ultrasound biomicroscopy: role in diagnosis and management in 130 consecutive patients evaluated for anterior segment tumours. R M Conway et al Br J Ophthalmol 2005;89:950–955

Study type Aims To demonstrate the use if UBM Included cysts and naevi in imaging anterior segment/ciliary body lesions Inclusion criteria Not stated Exclusion criteria Not stated Method Number of patients 130 (132 eyes) 45 melanomas where UBM and conventional ocular ultrasound was compared Outcomes Only 29% anatomical For anterior segment /CB correspondence between UBM melanoma only and conventional ultrasound

Conclusions UBM preferred for monitoring A/B scan (conventional US ) and measuring CB tumours. used for posterior melanoma and to detect posterior However A/B-scan can also be extension used to demonstrate features characteristic of melanoma including low internal reflectivity, internal acoustic hollowness and sound attenuation, choroidal excavation, orbital shadowing and the presence of spontaneous vascular pulsations and is therefore better than UBM at diagnosis (haemangioma verses melanoma)

Comment Superior to conventional ocular Largest series of patients ultrasound for CB and anterior investigated with UBM segment tumours esp good for serial observation Abstract

Background/aim: Ultrasound biomicroscopy (UBM) is an important tool for assessing anterior segment pathology. This study sought to evaluate UBM in the management of anterior segment tumours.

9 Methods: Retrospective analysis of medical records of consecutive patients referred to the ocular oncology unit, University of California San Francisco (UCSF), for suspected anterior segment tumours from 1999 to 2004.

Results: 132 eyes from 130 patients were evaluated, including 55 uveal melanomas (UM), 21 iris naevi, 30 iris cysts, and 26 remaining lesions. Of the melanomas, 45 were also evaluated with conventional A/B-scan. There was 29% correspondence between the anatomical structures invaded by melanoma as identified by B-scan v disease extent defined by UBM. Ciliary body and peripheral iris involvement by melanomas was significantly more fre- quently observed by UBM than B-scan. Seven of 30 benign cysts were diagnosed as cystic before UBM evaluation. In three cases, neuroepithelial cysts were associated with intercurrent pathology including iris naevus (n = 2) and ciliary body melanoma (n = 1). Two ciliary body melanomas showed cavitation, including one patient with a pseudocyst. Histopathological correlation was possible in six cases.

Conclusion: UBM is an indispensable tool for the management of anterior segment tumours. This study demonstrates the superiority of UBM v conventional B-scan for the precise localisation of uveal melanoma, especially involving the ciliary body and peripheral iris.

10 6.PET/CT imaging: detection of choroidal melanoma S Reddy et al Br J Ophthalmol 2005;89:1265–1269

Study type Prospective cases Aims To compare uvea melanoma dimensions using PET/CT, conventional ocular ultrasound and ophthalmoscopy Inclusion criteria SUV>2.5 considered positive Exclusion criteria Not detailed Method Comparative study at diagnosis Based on AJCC classification of tumour sizes T1, T2, T3 and COMS group classication Number of patients 50 Outcomes Smallest tumour physiologically identifiable by PET/CT had basal dimensions of 3x5.9 and an apical height of 2.9 mm.

No small tumours (T1) could be detected with PET/CT Conclusions PET/CT Comment Unlikely that PET/CT will be Imaging not purely able to distinguish between naevi dependent on tumour size, and melanoma functionally fused PET/CT was used but not all intraocular MM have SUV>2.5 some barely have any metabolic activity

Aim: To determine the size of untreated choroidal melanomas resolved by whole body positron emission tomography fused with computed tomography (PET/CT). Methods: 50 consecutive patients with untreated choroidal melanomas underwent whole body PET/CT. A functionally fused helical CT scan and 18-fluoro-2-deoxyglucose (FDG) PET scans were employed. The tumours were identified (both quantitatively and qualitatively) and compared with clinical measurements derived from ophthalmoscopic, angiographic, and ultrasonographic imaging. Standardised uptake values (SUV) of more than 2.5 were considered positive.

Results: Among the 50 patients with choroidal melanoma, PET/CT scan SUVs of more than 2.5 were noted in 14 (28%) tumours. No AJCC T1 class tumours, 33.3% of T2 melanomas, and 75% of T3 melanomas were physiologically identifiable on PET/CT. With respect to COMS group classifications, no small choroidal tumours, 33% of medium, and 75% of large melanomas were physiologically identifiable. The sole ring melanoma was identifiable on PET/CT imaging. The smallest tumour physiologically

11 identifiable by PET/CT had basal dimensions of 3x5.9 and an apical height of 2.9 mm.

Conclusion: Though PET/CT was found to be capable of physiologically identifying certain medium (T2) and most large sized (T3) choroidal melanomas, physiological imaging was not completely dependent upon tumour size. Functionally fused PET/CT localised the tumours within the eye and assessed their physiological activity.

12 Intraocular Biopsy

1.Iris ring melanoma: fine needle biopsy. D H Char et al

Br J Ophthalmol 2006;90:420–422

Study type Interventional case series Aims Describe FNAB for ring Very difficult diagnosis to melanoma make Very lethal subgroup of uveal melanoma Often present late having had glaucoma filtering operations Inclusion criteria Presumed or suspected ring MM Exclusion criteria Method Single centre Retrospective case not analysis Biopsy performed usings a transcornal route 1800 degress away from the tumour Number of patients 22 patients, 16 of which had FNAB to make the diagnosis Outcomes 11 of 16 biopsies were positive 5 false negatives paucicellular aspirate no cases of extra-scleral extension Conclusions FNAB can be used to make the None of the cases were diagnosis of ring melanoma detected with UBM or conventional ultrasound Comment Difficult to perform Difficult to report The author recommends an open biopsy when a false negative result is seen Abstract

Aims: To delineate the diagnostic accuracy of fine needle biopsy in iris ring melanoma and determine the tumour related mortality of this neoplasm. Methods: A retrospective analysis of 22 patients with iris melanomas that involve the entire 360 degrees of the anterior chamber angle.

Results: Iris ring melanomas were correctly diagnosed in all cases. In 11 of 16 cases (69%) a fine needle biopsy performed 180 degrees away from the main mass was positive for an iris ring melanoma. The tumour related mortality in iris ring melanoma cases was four of 22 patients (18%). Actuarial survival analysis showed a 10 year mortality (Kaplan-Meier) of 15%.Conclusion: A fine needle aspiration biopsy can be used to diagnose an iris ring melanoma. Iris ring melanomas have significant mortality compared with focal tumours.

13 2.Choroidal Biopsies for Intraocular Tumors of Indeterminate Origin. Kvanta et al AJO Dec 2005

Study type Retrospective, non comparative interventional case series Aims Determine the diagnosis of a choroidal mass Inclusion criteria Not stated Exclusion criteria Not stated but “case selection is important” Method 3 port pars planar vitrectomy with diamond knife incisional biopsy of the tumour post incision diathermy Number of patients 10 Outcomes A diagnosis was achieved in all Very high complication rate 10 patients. 5 had melanoma

Complications: Intraocular haemorrhage in 1 of Specimens fixed and paraffin 10 patients (one led to phitisis block sliced at 4 micrometre and loss of the eye) intervals 6 eyes removed 2 eye suffered retinal detachment

Conclusions Useful technique but high local complication rate Comment Due to high risk of complications the authors suggest a needle biopsy first Abstract (unable to cut and paste)

14 3. DIAGNOSTIC TRANSVITREAL FINE-NEEDLE ASPIRATION BIOPSY OF SMALL MELANOCYTIC CHOROIDAL TUMORS IN NEVUS VERSUS MELANOMA CATEGORY. James J. Augsburger et al. Trans Am Ophthalmol Soc 2002;100:225-234

Study type Aims To report the use of FNAB to determine the diagnosis of an indeterminate choroidal mass? Melanoma or naevus Inclusion criteria Suspicious choroidal naevi 1.5- 3mm thick ,10mm diameter Exclusion criteria Method Retrospective personal series of FNAB with 25G needle Number of patients 34 between 1983 and 2001 Outcomes The biopsy yielded a sufficient Follow wide range but aspirate for cytodiagnosis in 22 median is 2.6 years. Longer of 34 cases (64.7%) follow up would be better to prove FNAB was accurate Therefore in 12 (35%) there were insufficient cells to make the diagnosis. These patients were observed. 4 of 12 revealed subsequent growth and the lesion was reclassified as a melanoma

Conclusions Many of these lesions would What about sampling error? have been termed choroidal melanoma without diagnostic biopsy Comment

Abstract

Purpose: To report an experience with fine-needle aspiration biopsy of selected small melanocytic choroidal tumors during the interval from April 13, 1983, through January 19, 2001.

Methods: Retrospective descriptive case series report of 34 patients with a small melanocytic choroidal tumor (maximal diameter, mm; thickness, mm but mm) evaluated diagnostically by transvitreal fine-needle aspiration biopsy prior to treatment. None of the tumors had invasive features at the time of biopsy.Results: Patients ranged in age from 26 to 73 years (mean, 50.9 years). The evaluated choroidal tumors had a mean maximal basal diameter of 8.0 mm and a mean maximal thickness of 2.4 mm. Eighteen of the 34 tumors (52.9%) had been documented to enlarge prior to biopsy. Biopsy was performed in all cases using a 25-gauge hollow lumen needle and a transvitreal approach via a pars plana puncture site. The biopsy

15 yielded a sufficient aspirate for cytodiagnosis in 22 of 34 cases (64.7%). In these cases, the tumor was classified as malignant melanoma in 16 (47.1% of total), intermediate lesion in 4 (11.8%), and benign nevus in 2 (5.9%). The 12 tumors that yielded an insufficient aspirate and the four lesions that yielded intermediate cells continued to be classified as “nevus versus melanoma” and were monitored periodically for growth or other changes. Four of the 12 tumors that yielded an insufficient aspirate for cytodiagnosis and all four lesions that yielded intermediate cells were eventually reclassified as small choroidal melanomas and treated. The remaining eight tumors that yielded an insufficient aspirate and the two tumors that yielded benign nevus cells were classified as benign nevi at the most recent follow-up evaluation.Conclusions: Fine-needle aspiration biopsy showed that a substantial proportion of small melanocytic choroidal tumors likely to be classified clinically as small choroidal melanomas in many centers were in fact benign nevi or intermediate lesions.

16 4. Transvitreal fine needle aspiration biopsy: the influence of intraocular lesion size on diagnostic biopsy result. Cohen et al. Eye (2001) 15, 143-147

Study type Retrospective case note analysis Aims To determine the Inclusion criteria Indeterminate choroidal masses which underwent pars planar transvitreal needle biopsy from 1986-1999 Exclusion criteria Method All tumour heights were recorded with A scan ultrasonography. Number of patients 83 Outcomes There was insufficient material A yield of 88% for cytological examination in 10 cases, and sufficient material in 73 cases (an overall diagnostic report rate was 88%). Yield 96% accurate result increased significantly with increasing tumour height. Only 1 false positive FNAB showed melanoma but <2= 40% 2-4mm = 90% histopathology report was >4 =98% prostate met

FNAB result correlated with pathology of enucleated specimen in 26 of 27 cases No RD (96%) No Seeding Transient vitreous haem in 24%

Conclusions Comment This paper helps guide patients Outcome is surgeon and and surgeons to select which size pathologist specific of tumour to biopsy Abstract

Purpose To detennine the efficacy of transvitreal biopsy in the diagnosis of suspected intraocular malignancy and simulating conditions.

Methods We perfonned a retrospective study of the case notes from patients who underwent pars plana transvitreal biopsy from July 1986 to October 1999. We studied the relationship between lesion thickness as measured by A• scan ocular ultrasound and the incidence of a successful diagnostic biopsy. We assessed the diagnostic accuracy by comparing the biopsy result with the histological examination of any subsequently enucleation specimens and noted the incidence and severity of complications attributable to the biopsy. Results A total of 83 biopsies were performed

17 for choroidal masses. There was insufficient material for cytological examination in 10 cases, and sufficient material in 73 cases (an overall diagnostic report rate was 88%). There was a strong correlation (p = 0.0004, Mann-Whitney U-test) between a diagnostic biopsy result and the thickness of the lesion on A-scan ultrasound: a biopsy was diagnostic in only 40% (4 of 10) of choroidal lesions less than 1.99 mm thick, whereas biopsies taken from lesions between 2.00 and 4.00 mm thick were diagnostic in 90% of cases (27 of 30). In thicker lesions of 4 mm or more the cell aspirate was sufficient to make a diagnosis in 98% (42 of 43). Following diagnostic biopsy 27 patients had their tumours resected, and the histology results following enucleation confined the cytological diagnosis of malignancy in 96% of these cases (26 of 27). Conclusion Transvitreal biopsy is a highly accurate diagnostic procedure with a low complication rate. It is a reliable diagnostic tool in suspicious choroidal lesions greater than 2 mm thick.

18 Sensitivity and Specificity of Ultrasonography, Fluorescein Videoangiography, Indocyanine Green Videoangiography, Magnetic Resonance and Radioimmunoscintigraphy in the Diagnosis of Primary Choroidal Malignant Melanoma. Romani et al Ophthalmologica 1998;212(suppl 1):44–46

Study type Comparative small case series Aims To determine the best diagnostic test for choroidal melanoma Inclusion criteria Not stated Exclusion criteria Not stated Method Diagnosis compared against histology in 4 cases and against presumed diagnosis in the rest

Number of patients 12 (9 melanomas) Very small number for sensitivity specificity study Outcomes Radioimmunoscintigraphy FV and ICFV present a high (sensitivity>67%) was less sensitivity (100%) and good sensitive than ultrasonography, specificity (82%), but these MR and angiography. examinations cannot be used when the lesion presents a Specificity was good for all the pre-equatorial localization or considered examinations; it was if opacities of dioptric media 92% for ultrasonography, 87% are present for RIS, 83% for MR and 82% for angiography. MR cannot present a very high specificity; in fact, other bulbar lesions contain paramagnetic substances with a signal similar to melanin

Specificity of MR was 83%, because of the presence of methemoglobin, a paramagnetic sub- stance, in 2 lesions which were false- positive. The specificity of MR was very high (100%).

Conclusions In our series of cases ultrasonography is a highly reliable diagnostic technique: sensitivity>100%, specificity>92%.

Comment Conventional Ocular Ultrasound (A/B scan) is the best test to diagnose choroidal melanoma

19 Abstract

The prognosis of primary choroidal malignant melanoma (PCMM) is fatal if no early reliable diagnosis is performed. Any incisional biopsy is impossible, and diagnosis is only based on instrumental examinations. The purpose of this study is to evaluate the reliability of ultrasonography, fluorescein videoangiography (FV), indocyanine green videoangiography (ICGV), magnetic resonance (MR) and radioimmunoscintigraphy (RIS) in the diagnosis of PCMM in a series of 12 eyes in which the tumor was suspected. A presumed diagnosis of PCMM was made when a positive result was obtained with 3 or more methods. The presumed diagnosis was then compared with histological findings

(true value) in 4 enucleated eyes. The sensitivity and specificity of every single method were evaluated comparing its results with the final presumed diagnosis and with the histological findings. Sensitivity and specificity of every single method have been expressed as percentage of correspondence with the presumed diagnosis. Sensitivity was 100% for ultrasonography, MR, FV, ICGV and 67% for RIS. Specificity was 92% for ultrasonography, 87% for RIS, 83% for MR and 82% for FV and ICGV. This study indicates that the ophthalmologist can obtain a good diagnostic reliability in the case of PCMM using only ultrasonography, FV and ICGV. Besides MR and RIS are important adjunctive methods to ophthalmological investigations for the diagnosis of PCMM.