UVEAL MELANOMA: EPIDEMIOLOGY, ETIOLOGY, CLUSTER DETECTION, AND CASE FOR A RARE TUMOR REGISTRY Marlana Orloff, MD Assistant Professor Department of Medical Oncology Thomas Jefferson University Hospital
UVEAL MELANOMA: EPIDEMIOLOGY, ETIOLOGY, CLUSTER DETECTION, AND CASE FOR A RARE TUMOR REGISTRY
Marlana Orloff, MDAssistant ProfessorDepartment of Medical OncologyThomas Jefferson University Hospital
FINANCIAL DISCLOSURE Nothing to disclose
WHAT WE KNOW Rare cancer Most common primary intraocular tumor Incidence: 7 per million per year in US Approximately 2000 new cases diagnosed
annually in the US Median age 60yrs Overall Survival varied *Above based on what is reported to
SEER/CINA national databases / single center experiences
WHAT WE KNOW More common in lighter eyed people White:Non-White 175:1 M:F slight M>F Very rarely hereditary
BAP1 Syndrome BRCA2
WHAT WE DON’T KNOW What causes it?
PREVIOUS EPIDEMIOLOGIC STUDIES Singh et al 2012
Queried SEER database 1973-2008 4,070 cases
Only included cases of the “uveal tract” Excluded cases of site unspecified
Concern if 2000 cases to be diagnosed annually in the US then this is a gross underestimation of number of cases
ATTEMPT TO SEE TRENDS IN DIAGNOSIS AND CAUSE
Eye cancer incidence 2006-2010 taken from North American Association of Central Cancer Registries (NAACCR)
Assumed ocular melanoma to be majority of cases Data from Washington, DC was censored due to
small numbers (N=9) Lowest rates in South Dakota
0.29/100,000 (13 cases) Highest rates in Oregon
1.23/100,000 (242 cases)
UNUSUAL CLUSTERING
PREVIOUS DESCRIBED CLUSTER: 1982
Patient Gender
Age at Diagnosi
s
Date of Diagnosi
s
Residence at
Diagnosis
Institution that made
Official Diagnosis
Residence at Time of Suspected Clustering
Dates at least part time at Residence
During Time of Suspected
Clustering
1 F 22 4/13/2009 Aiken, SC
Wills EyePhiladelphia,
PAHuntersville
, NC 2002-2005
2 F 24 12/2008 Charleston, SC
Duke Eye Center
Durham, NCHuntersville
, NC 1994-2007
3 F 20 7/8/2013 Charlotte, NC
Wills EyePhiladelphia,
PAHuntersville
, NC 2006-2011
4 F 31 2/7/11 Charlotte, NC
Wills Eye Philadelphia,
PAHuntersville
, NC2005/2007-
2013
5 F 30 2/2014 Concord, NC Duke Eye
CenterDurham, NC
Huntersville, NC 2005-2007
Patient Gender
Age at Diagnosi
s
Date of Diagnosi
s
Residence at
Diagnosis
Institution that made
Official Diagnosi
s
Residence at
Time of Suspect
ed Clusterin
g
Dates at least part time at Residence
During Time of Suspected
Clustering
1 F 31 10/1/2001
Birmingham, AL
Wills Eye Philadelphia, PA
Auburn, AL
(Auburn University
)
1989-1993
2 F 31 1/2000 Birmingham, AL
Eye Foundatio
nBirmingham, AL
Auburn, AL
(Auburn University
)
1989-1993
3 F 42 7/2012 Memphis, TN
Hamilton Eye
InstituteMemphis,
TN
Auburn, AL
(Auburn University
)
1989-1993
OUR INVESTIGATION Contact individual state registries to look at
cases reported Calculation of STANDARD INCIDENCE RATIO
(SIR)
>1 suggests observed > expected NC (Mecklenberg): 0.7 (0.5-0.9) AL (Statewide only Females): 1.15 (0.98-1.32)
Lower limit =)Expected#(x2
2
/2 observed),x(#2 Upper Limit =)Expected#(x2
2
/2-1 1),observedx(#2
Where
2
,vis the 100α percentile of the chi-square distribution with v degrees of freedom and α is 0.05.
OUR INVESTIGATION: ROAD BLOCKS Issues with case reporting: Accuracy, Timing, Too Few Numbers
“The incidence rates may be underestimated because the N.C. CCR (North Carolina Central Cancer Registry) does not have complete cases for 2012 and 2013”.
Further, they noted that they had not had any data for cases diagnosed in 2014 due to lag time in reporting.
In addition they also noted in their report that “cases diagnosed out of the state and country but that may be receiving treatment in facilities in North Carolina were not included as they are not required to be reported to the N.C. CCR”.
When one looks at the details regarding the individual cases in the cluster, NONE of which would have been expected to be included as observed cases in the final analysis One diagnosed in 2014 Two lived in SC at the time of diagnosis Two though living in NC at the time of diagnosis were diagnosed in PA
When the state was asked to provide a report that would break down the observed and expected incidences based on age and gender, such data was not able to be generated due to “small numbers of observed cases”.
OTHER ISSUES WITH CASE REPORTING As of October 2016, a concise ICD-10 code
for “uveal melanoma” does not exist Rather it is coded as “choroidal/ciliary body/iris
neoplasm” or “melanoma of the face unspecified”
Often no biopsy for histologic diagnosis Often path report part of report to cancer
registry Current reporting of cases does not include
much beyond diagnosis
INVESTIGATING CANCER CLUSTERSREVIEW OF LAST 20 YEARS
EVALUATING CANCER CLUSTERS IN 2016 “Rapid Case Ascertainment” “Reconstruction Residential Histories” “Application of Spatial Analytics” “Continuous Monitoring of Registry Data”
PROPOSAL FOR REGISTRY National US UM Registry Cases directly entered by participating sites
Limited diagnosing and treating centers Allow for:
Accurate and quick case reporting Real time incidence and clustering surveillance Prospective and retrospective investigation
SAMPLE DATA SETTable 3. Sample Data Set
Demographicsa Tumor Genetics Occupation History
Medication Use
Host Factorsb Details of Primary Treatment
Hobbiese Radiation Exposure/Electromagnetic Exposuresh
Lifetime Residential History
Adjuvant Treatment
History of other Cancersf
Surveillance Schedule*
Tumor Characteristics
Clinical Trial Participation
Family Historyg
UV Exposurei
Pre-existing Nevic
Relapse Date Date and Cause of Death
Other Toxic or Environmental Exposuresj
Metastases Characteristics
Previous Ocular Historyd
Histologic Information
Other Medical Historyka: Age, gender, ethnicity b: Eye color, skin color, hair color c: Choroidal nevi d: Trauma, uveitis, infection e: Welding, cooking, other f: Melanoma, mesothelioma, renal cell carcinoma, meningioma, breast cancer, pancreatic cancer g: Melanoma, mesothelioma, renal cell carcinoma, meningioma, breast cancer, pancreatic cancer h: electromagnetic fields [19] i: Sunlight, blue light, black light, tanning bed use i: Pesticides, PCB, phenols j: Prior atypical fungal infections, charcot-marie tooth, valvular disorders
GEOGRAPHIC SURVEILLANCEEXAMPLE: 10 CASES DIAGNOSED IN 2016
2003
2008
GEOGRAPHIC SURVEILLANCE10 CASES DIAGNOSED IN 2016
2008
ARCGIS: SPATIAL SOFTWAREDetecting and Quantifying Patterns
Analyze your data to reveal patterns or show areas of concentrations. Find statistically
significant spatial clusters of high values (hot spots) and low values (cold spots) in your data.
CONCLUSIONS Uveal melanoma is a rare cancer and may be difficult to
accurately report to state and national cancer registries We don’t know what causes it Efforts need to be focused towards the establishment of a
rare cancer registry for UM Accurate case reporting Real time geographic surveillance Can then lead to appropriate testing of hypotheses
Implementing spatial analytics to a rare cancer registry for the purposes of geographic pattern detection and cluster surveillance to our knowledge will be the first of its kind
Noticeable trends may then lead to the first ever opportunities for primary and secondary prevention
A properly designed rare cancer registry for UM may also encourage the development of such databases for other rare cancers
THANK YOU