Informe de Proyecto de Doctorado en Operativa Estad stica e Investigacion

Statistical methods of small area estimation for the study of cancer morbility in Small Animals

Daniel Mart nez Bello, dadymar@correo.uv.es

Universitat de Valencia

June 19, 2013

Informe de Proyecto de Doctorado en Operativa Estad stica e Investigacion

Statistical methods of small area estimation for the study of cancer morbility in Small Animals

Daniel Mart nez Bello, dadymar@correo.uv.es

Universitat de Valencia

June 19, 2013

Table of contents

Motivating study

Analysis of retrospective small animals cancer cases based in hospital records

Cancer in small animals

What is cancer?
Neoplasms are abnormal new growths of tissue which develop faster than adjacent normal tissues and in an uncoordinated, persistent manner. benign or malignant cancer is generally restricted to the malignant growths. Uncontrolled proliferation. Impaired cellular differentiation Altered cell communication and adhesion.[Morris and Dobson, 2001]

Cancer in small animals

Causes of cancer
Spontaneous genetic events. External stimuli: biological (viruses, parasites, hormones) physical (UV light, radiation, trauma) chemical Inherited genetic events (familial cancers).

Differences between benign and malign neoplasms [Morris and Dobson, 2001]
Benign Relatively slow Growth may cease in some cases Expansive Usually well dened boundary between neoplastic and normal tissues. May become encapsulated. Often minimal May cause pressure necrosis and anatomical deformity Malignant Often rapid Rarely ceases growing Invasive Poorly dened borders, tumour cells extend into and may be scattered throughout adjacent normal tissues. Often serious Tumour growth and invasion results in destruction of adjacent normal tissues, manifest as ulceration of supercial tissues, lysis of bone Occurs by lymphatic and haematogenous routes and transcoelomic spread Often life-threatening by virtue of destructive nature of growth and metastatic dissemination to other, vital organs.

Does not occur

Often minimal (can be life threatening if tumour develops in a vital organ, e.g. brain)

Tumour classication by tissue of Origin: Mesenchymal or Epithelial

Tissues and the tumors derived from them are divided into mesenchymal or epithelial origin. Mesenchymal elements include connective tissue, striated and smooth muscle, blood cells, and endothelial cells and related tissues (synovium, mesothelium, and meninges). Epithelial cells include squamous epithelia of the skin, cells that line the respiratory, digestive, urinary, and reproductive tracts, all glands, exocrine and endocrine, and cells of neuroectoderm origin such as melanocytes.

Tumour classication by tissue of origin: Mesenchymal

The tissue of origin and the sufx -oma designate benign mesenchymal neoplasms (e.g. broma). The tissue designation and the sufx -sarcoma designate malignant neoplasms of mesenchymal(e.g. brosarcoma).

Tumour classication by tissue of origin: Epithelial

the tissue of origin with the sufx -adenoma designates benign epithelial neoplasms of glandular origin (e.g. mammary adenoma). Benign epithelial neoplasms that arise from lining epithelium are usually termed papillomas. The tissue of origin and the sufx -carcinoma is used for malignant epithelial neoplasms. Those that make histologically evident glands within the neoplasm are termed adenocarcinomas. Exceptions melanoma and hepatoma, which refer to malignant neoplasms (malignant melanoma and hepatocellular carcinoma are more accurate). Lymphoma, despite the above objections, is the preferred term.

Cancer in small animals

Why do we study cancer in small animals?
The canine lifespan is around 1/7 of the human being. There are many different phenotypic breeds. Pure breeds represent an important model to study the importance of inheritance and environment in development and progression of cancer (and not only this disease). Incidence of cancer in dogs may be compared to that in humans, with differences in relative risk of various types of cancer. Dogs and man share the same environmental effects. Knowing more about the epidemiological aspects of the disease allow prevention measures and treatment. Spontaneous cancer offers a possibility to make cancer research without doing experimentation [Gamlem et al., 2008].

Communes from Bucaramanga, 17 administrative divisions

Comunas en Bucaramanga
There are 17 commune in the city of Bucaramanga.

Communes from Bucaramanga, 17 administrative divisions

Frequency Proportion 2008 22 0.06 2009 8 0.02 2010 81 0.24 2011 106 0.31 2012 126 0.37 Total 372 1.00

Total number of cases included in the city of Bucaramanga, with some observations outside of the city. 2008 16 0.09 2009 6 0.04 2010 31 0.18 2011 46 0.27 2012 70 0.41 Total 170 1.00

Frecuency Proportion

Cancer cases excluding cases from Sticker tumor. 2008 14 0.09 2009 6 0.04 2010 28 0.19 2011 38 0.25 2012 65 0.43 Total 152 1.00

Frecuency Proportion

Communes from Bucaramanga, 17 administrative divisions

7.150

G G G G G G G G G G G GG G G G G G GG G G G G G G G

GG

7.125

G G G GG

G G G

G G G

G G G G G G G G G G G
7.100

G G G G

G G G G G G G G G G G G G GG G G G GG G G G G G GG G G GG GG G G G G G G G G G GG G G G G G GG G G G G G G GG G G G G G G G G G G G G G G G G

Ao
G G G G G

2008 2009 2010 2011 2012

lat

G 2

G G G G G

G G G G

G G

GG G G

GG

7.075

73.14

73.12

73.10

long

Cancer cases in Bucaramanga for 2008 and 2009

2008 2009

7.150

7.150

G G

7.125

7.125
GG

lat

G G G G G

lat

G G

7.100

7.100
G

G G G

7.075

7.075

73.14

73.12

73.10

73.14

73.12

73.10

long

long

Cancer cases in Bucaramanga for 2010 and 2011

2010 2011

7.150

7.150
G

G G G G G G G G G G G G G G G

G G

G G G G

G G G G

7.125

G G G G

7.125
G G G

G G

lat

lat

G G G G

G G G G

G G

G G G

G G G G G

7.100

G G

7.100

G G

7.075

7.075

73.14

73.12

73.10

73.14

73.12

73.10

long

long

Cancer cases in Bucaramanga for 2012

2012

7.150

G G

G G G G G G G G G G G GG G GG G G GG G G G G G G G G G G GG G G G G G G G G G G G G

7.125

lat

G G G G G G G G G G

7.100
G G G G G G G G

7.075

73.14

73.12

73.10

long

Cancer cases distribution by diagnostic type

2008 2009 2010 2011 2012

Clin 13 5 17 30 40 105

Spatially located cases Count Proportion Hist Total Clin Hist 1 14 0.93 0.07 1 6 0.83 0.17 10 27 0.63 0.37 8 38 0.79 0.21 25 65 0.62 0.38 45 150 0.70 0.30

Total cases Clin 21 7 22 63 60 173 Count Hist 1 1 59 43 66 170 Total 22 8 81 106 126 343 Proportion Clin Hist 0.95 0.05 0.88 0.12 0.27 0.73 0.59 0.41 0.48 0.52 0.50 0.50

Table by sex Spatially located cases Count Proportion M Total F M 3 14 0.79 0.21 2 6 0.67 0.33 13 27 0.52 0.48 14 38 0.63 0.37 19 65 0.71 0.29 51 150 0.66 0.34 Total cases F 17 6 44 67 82 216 Count M NR 5 0 2 0 35 2 39 0 44 0 125 2 Total 22 8 81 106 126 343 F 0.77 0.75 0.54 0.63 0.65 0.63 Proportion M 0.23 0.25 0.43 0.37 0.35 0.36 NR 0.00 0.00 0.02 0.00 0.00 0.01

2008 2009 2010 2011 2012

F 11 4 14 24 46 99

Cancer cases distribution by body system

Total cancer cases by distribution in body system.
08 1 1 13 0 2 0 2 2 0 1 0 22 09 0 0 7 0 0 0 0 0 0 1 0 8 10 3 3 20 4 7 0 25 12 0 1 6 81 Count 11 6 7 52 4 12 0 14 5 1 0 5 106 12 2 5 59 6 27 1 15 6 0 2 3 126 Total 12 16 151 14 48 1 56 25 1 5 14 343 08 0.08 0.06 0.09 0 0.04 0 0.04 0.08 0 0.20 0 0.06 09 0 0 0.05 0 0 0 0 0 0 0.20 0 0.02 Proportion 10 0.25 0.19 0.13 0.29 0.15 0 0.45 0.48 0 0.20 0.43 0.24 11 0.50 0.44 0.34 0.29 0.25 0 0.25 0.20 1 0 0.36 0.31 12 0.17 0.31 0.39 0.43 0.56 1 0.27 0.24 0 0.40 0.21 0.37

Digestive Generalized Genital Hematopoyetic Muscular-Esqueletic Muscular-Esqueletic-Urinary Sense organs Skin-anexa Respiratory Urinary No Register

Digestive Generalized Genital Hematopoyetic Muscular-Esqueletic Muscular-Esqueletic-Urinary Senses organs Skin-anexa Respiratory Urinary No Register

08 1 1 7 0 1 0 2 1 0 1 0 14

09 0 0 5 0 0 0 0 0 0 1 0 6

10 1 1 6 2 1 0 7 6 0 0 3 27

11 2 3 16 2 5 0 3 3 1 0 3 38

12 0 3 31 1 15 1 11 1 0 1 1 65

Total 4 8 65 5 22 1 23 11 1 3 7 150

08 0.25 0.12 0.11 0 0.05 0 0.09 0.09 0 0.33 0 0.09

09 0 0 0.08 0 0 0 0 0 0 0.33 0 0.04

10 0.25 0.12 0.09 0.40 0.05 0 0.30 0.55 0 0 0.43 0.18

11 0.50 0.38 0.25 0.40 0.23 0 0.13 0.27 1 0 0.43 0.25

12 0 0.38 0.48 0.20 0.68 1 0.48 0.09 0 0.33 0.14 0.43

Cancer cases distribution by category

Total cancer cases by category.

2008 2009 2010 2011 2012 Benigno 18 7 36 69 76 206 Maligno 3 1 44 33 47 128 No Registra 1 0 1 4 3 9 Total 22 8 81 106 126 343 Benigno 0.82 0.88 0.44 0.65 0.60 0.60 Maligno 0.14 0.12 0.54 0.31 0.37 0.37 No Registra 0.05 0 0.01 0.04 0.02 0.03

2008 2009 2010 2011 2012

Benigno 10 5 17 27 41 100

Maligno 3 1 9 10 21 44

No Registra 1 0 1 1 3 6

Total 14 6 27 38 65 150

Benigno 0.71 0.83 0.63 0.71 0.63 0.67

Maligno 0.21 0.17 0.33 0.26 0.32 0.29

No Registra 0.07 0 0.04 0.03 0.05 0.04

Table : Anatomic location of total cancer cases

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 Location Glandula Mamaria Vulva Boca Generalizado No Registra MA Pene Cuello Ojo Perianal MP Testiculo Abdomen Bazo Dorso Encia Fosa Nasal Higado Piel Torax - Pecho Vejiga Cola Mandibula Oreja Parpado Tren Posterior Ganglio Popliteo Lengua Count 108 24 16 16 15 12 12 10 9 9 8 8 7 7 7 7 7 7 7 7 6 5 5 5 5 5 4 4 % 29.03 6.45 4.30 4.30 4.03 3.23 3.23 2.69 2.42 2.42 2.15 2.15 1.88 1.88 1.88 1.88 1.88 1.88 1.88 1.88 1.61 1.34 1.34 1.34 1.34 1.34 1.08 1.08 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 28 Location Prepusio Labio Arco Cigomatico Codo Corvejon Escroto Espacio Inter Digital Falange Ganglio Retromandibular Garganta Hueso Inguinal MA - Torax Mediastino MP - Rion Mucosa Oral Paladar piel Piloro Prostata Pulpejo Recto Retro Vulvar Seno Frontal Tercer Parpado Tonsilas Vesicula Count 3 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 NA % 0.81 0.54 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 NA

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

Location Glandula Mamaria Boca Generalizado No Registra Encia MP Cuello MA Torax - Pecho Bazo Cola Dorso Fosa Nasal Higado Ojo Perianal Tren Posterior Vejiga Abdomen Lengua Pene Piel Testiculo Ganglio Popliteo Garganta MA - Torax Mandibula MP - Rion Parpado Piloro Prepucio Vulva

Count 58 9 8 8 5 5 4 4 4 3 3 3 3 3 3 3 3 3 2 2 2 2 2 1 1 1 1 1 1 1 1 1

% 38.41 5.96 5.30 5.30 3.31 3.31 2.65 2.65 2.65 1.99 1.99 1.99 1.99 1.99 1.99 1.99 1.99 1.99 1.32 1.32 1.32 1.32 1.32 0.66 0.66 0.66 0.66 0.66 0.66 0.66 0.66 0.66

Table : Final diagnosis for the total cancer cases, 1 to 30 histopathological and clinical classication
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Diagnosis Adenoma Mamario Tumor Venereo Transmisible Hemangiosarcoma Carcinoma Mixto Mamario Carcinoma Escamocelular Papiloma Cutaneo No Registra Adenoma Glandula Cebacea Adenoma Glandulas Hepatoides Lipoma Carcinoma Mamario Leiomioma Linfoma Eplis Fibromatoso Carcinoma Celulas Transicionales Fibroma Hemangioma Lipoma Cutaneo Carcinoma Carcinoma Celulas Basales Carcinoma Oral Hemangioma Capilar Hemangioma Cutaneo Hemangiosarcoma Multicentrico Histiocitoma Linfosarcoma Mastocitoma Cutaneo Melanoma Osteosarcoma Plasmocitoma Cutaneo Count 74 30 18 14 12 11 9 8 8 8 7 7 7 6 4 4 4 4 3 3 3 3 3 3 3 3 3 3 3 3 % 19.89 8.06 4.84 3.76 3.23 2.96 2.42 2.15 2.15 2.15 1.88 1.88 1.88 1.61 1.08 1.08 1.08 1.08 0.81 0.81 0.81 0.81 0.81 0.81 0.81 0.81 0.81 0.81 0.81 0.81

Table : Final diagnosis for the total cancer cases, 31 to 60 histopathological and clinical classication
31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Diagnosis Tumor Celulas de Certoli Tumor Maligno Inltrante mal Diferenciado Adenocarcinoma Indiferenciado Adenoma tipo I Carcinoma Baso escamoso Carcinoma Escamoso de Celulas grandes Carcinoma Hepatico Carcinoma Indiferenciado Carcinoma Tiroide de Celulas Peq Hemangiosarcoma Cavernoso Histiocitoma Cutaneo Leiomiosarcoma Linfoma Cutaneo Linfosarcoma Mesenterico Mastocitoma Grado II diferenciacion Neurobroma Papiloma Escamoso Sarcoma indiferenciada Adenocarcinoma Inltrante Adenocarcinoma Moderadamente Diferenciado Adenoma Complejo Adenoma de Parpado Carcinoma Ductual con Diferenciacion Neuroendocrina Carcinoma Escrotal Carcinoma Glandula Cebacea carcinoma Glandulas Cebaceas inltrante Carcinoma Glandulas Tarsales Carcinoma Intestinal Carcinoma Papilar Inltrante Count 3 3 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1 1 1 % 0.81 0.81 0.54 0.54 0.54 0.54 0.54 0.54 0.54 0.54 0.54 0.54 0.54 0.54 0.54 0.54 0.54 0.54 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27

Table : Final diagnosis for the total cancer cases, 61 to 90 histopathological and clinical classication
61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 Diagnosis Carcinoma Solido Mamario Carcinoma Tubulo Papilar Mamario Carcioma Urotelial Papilar Inltrante Epitelioma Glandulas Hepatoides Fibrosarcoma Hemagioma Inltrativo Cavernoso Hemangioma Cavernoso Hemangioma Hepatico Hemangioma Palpebral Hemangiopericitoma Hemangiosarcoma Cutaneo Hemangiosarcoma Esplenico Hemangiosarcoma Ocular Histiocitoma Oral Linfoma Bucal Linfoma Celulas Mixtas Linfoma Cutaneo Inltrativo Linfoma de celulas Pequeas Linfoma Hepatico Linfoma Intestinal Linfoma Metastasico Progresivo Linfoma Nodal Linfoma Oral Linfosarcona Mastocitoma Cutaneo Inltrativo Mastocitoma Grado I diferenciacion Melanoma Corneal Melanoma Cutaneo Melanoma Inltrativo Melanoma Maligno Inltrativo Total Count 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 151 % 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 100

Table : Final diagnosis for the total cancer cases, 91 to 111 histopathological and clinical classication
91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 Diagnosis Melanoma oral Melanoma Palebral Melanoma Travecular Cutaneo Metastasis Mistiocitosis Cutanea Mixoma Cutaneo Osteoma Papiloma Oral Plasmocitoma Grado II diferenciacion Plasmocitoma Indiferenciado Rabdomioma Sarcoma Oral Sarcoma Sinovial Seminoma Solido Tumor Abdominal Multifocal Tumor Bifasico Tumor Celulas Basales Tumor Celulas Certoli Tumor Fusocelular Tumor Maligno Inltrativo Tumor Retrovulvar Count 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 % 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27 0.27

Table : Final diagnosis for the cancer cases located in Bucaramanga, 1 to 25 histopathological and clinical classication
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 Diagnosis Adenoma Mamario Papiloma Cutaneo Hemangiosarcoma No Registra Adenoma Glandula Cebacea Adenoma Glandulas Hepatoides Carcinoma Escamocelular Carcinoma Mixto Mamario Eplis Fibromatoso Lipoma Carcinoma Celulas Transicionales Carcinoma Oral Linfoma Carcinoma Carcinoma Mamario Fibroma Hemangioma Hemangioma Capilar Histiocitoma Linfoma Cutaneo Linfosarcoma Lipoma Cutaneo Adenoma tipo I Carcinoma Celulas Basales Carcinoma Escrotal Count 52 9 6 6 5 5 4 4 4 4 3 3 3 2 2 2 2 2 2 2 2 2 1 1 1 % 34.44 5.96 3.97 3.97 3.31 3.31 2.65 2.65 2.65 2.65 1.99 1.99 1.99 1.32 1.32 1.32 1.32 1.32 1.32 1.32 1.32 1.32 0.66 0.66 0.66

Table : Final diagnosis for the cancer cases located in Bucaramanga, 26 to 47 histopathological and clinical classication
26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 Diagnosis Carcinoma Hepatico Carcinoma Tiroide de Celulas Peq Hemangiosarcoma Cavernoso Histiocitoma Cutaneo Leiomioma Leiomiosarcoma Linfoma Bucal Linfoma de celulas Pequeas Linfoma Intestinal Linfoma Metastasico Progresivo Mastocitoma Cutaneo Mastocitoma Grado II diferenciacion Melanoma Corneal Melanoma Cutaneo Melanoma Palebral Metastasis Papiloma Escamoso Tumor Abdominal Multifocal Tumor Celulas Certoli Tumor Celulas de Certoli Tumor Maligno Inltrativo Tumor Retrovulvar Count 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 % 0.66 0.66 0.66 0.66 0.66 0.66 0.66 0.66 0.66 0.66 0.66 0.66 0.66 0.66 0.66 0.66 0.66 0.66 0.66 0.66 0.66 0.66

Study objectives

General objectives
Developing methods for relative risk estimation of the most common neoplasms diagnosed in the city, using veterinary clinics and hospital registers.

Study objectives
Specic objectives
To dene a survey procedure (face to face, telephonic, internet, ?) to estimate canine (and feline?) population at neighborhood level, by sex, age and breed. To develop a simulation methodology for testing the proposed survey procedure. To develop the survey procedure in the geographic area of Bucaramanga. To establish and collect auxiliary information at area level, for small area estimation of canine (and feline?) neoplasms, by checking information sources from public (city hall, Planeation ofce)and private origin ().

Study objectives
Specic objectives
If the census information is available, the canine (and feline?) population estimation by combining the census information and the survey information, although this possibility is not a primary interest. Combining the information of veterinary clinic and hospital registers and the survey data, estimating the type and number of neoplasms in the canine (and feline?) population. Developing methods for relative risk estimation of the most common neoplasms diagnosed in the city, using veterinary clinics and hospital registers, using the spatial characteristics of the problem.

Estimation of population characteristics at small area level

Direct estimation
Horvitz-Thompson estimator for the total and the mean Ydirect =
i s

1 yi i

Y direct =
i s

1 i yi 1 i s i n N

Where i is the inverse probability, and equals pi =

Estimation of population characteristics at small area level

Generalised regression estimator

GREG estimator, based in linear regression. N 1 yj + k YGREG = xp j
j s k p=1

j s

1 xj j

Estimation of population characteristics at small area level

Model Based, Fay and Herrington estimator

Combination of direct estimations with linear regression, for areal level. 2 i = i + i Y i N (0, i ) i i = + X i is a direct estimator, 2 is a design varince, X is a Where Y i vector of area level covariates, and and are estimated by means of generalized least squares.

Estimation of population characteristics at small area level

Synthetic estimator

Estimation of population characteristics at small area level

Composite estimator
Combination of properties from direct and model based estimators d ,i and model based estimators Using direct estimators Y m,i , a combined estimator Y c ,i is determined by Y c ,i = i Y d ,i + (1 i )Y m ,i Y Where i is a shrinkage parameter to weight both estimators.

Bibliography

Gamlem, H., Nordstoga, K., and Glattre, E. (2008). Canine neoplasiaintroductory paper. APMIS. Supplementum, (125):518. Morris, J. and Dobson, J. (2001). Small Animal Oncology. Blackwell Science Ltd, Oxford, UK.