Ultrasonography of Osteosarcoma in Dogs
Sonography Atlas of Dogs

For each breed and group, the prevalence was estimated based on (sometimes small) samples. 95% confidence intervals were obtained using the exact binomial distribution.

Pmaxis the maximal prevalence for which a number of cases equal or inferior to the actual number of OSA cases by breed or by group was observed with a probability equal to 2.5%.

Pmin is the minimal prevalence for which a number of cases equal or superior to the actual number of OSA cases by breed or by group was observed with a probability equal to 2.5%.

Pmax =

Pmin =

r: number of cases observed

n: number of dogs that belong to a breed or a group

B( i |n, Pmax) and B(i |n, Pmin): probability to observe a “i” event in the binomial distribution and “n” is the repetition of events having individually a Pmax or Pmin probability.

If the confidence intervals between two breeds of two groups do not overlap, a significant difference in the OSA prevalence between the two breeds or groups is probably observed. For each breed and group, the number of breeds or groups not overlapping their confidence interval can be counted. Using these counts, ranking of the most frequently affected breeds and groups could be obtained.

 

The relation between the average dog heights and the OSA prevalence was reviewed because OSA prevalence could differ for larger animals with larger bones. This can be tested using a logistic regression of the disease status on the dog size. In addition, average heights were noted for both male and female dogs.

 

In total, 5373 dogs from 64 breeds were presented for surgical affections. Only 26 different breeds (Rottweiler (n= 14); Labrador Retriever (n=7); Belgian Shepherd; Golden Retriever, Dobermann, Beauceron (n=4 each); Bernese Mountain Dog, Pyrenean Mountain Dog, German Shepherd (n=3 each); Leonberger, Irish Wolfhound, Great Dane, Chow- Chow (n=2 each); Shar Pei, Scottish Terrier, Poodle, American Staffordshire Terrier, Bull Terrier, Irish Terrier, Brie Shepherd, Basset Fauve de Bretagne, Bordeaux Mastiff, Cavalier King Charles Spaniel, Lhasa Apso, Greyhound and Basset Hound (n=1 each) were retained, and a total of 67 cases of OSA were included in the study (Table 1).

The study included 39 cases of appendicular OSA (8 proximal humerus, 7 distal femur, 7 distal tibia, 5 proximal tibia, 4 distal radius, 3 olecranon, 2 distal ulna, 1 scapula, 1 ischium, 1 metacarpus), 12 cases of axial OSA (5 mandibular, 5 maxillary, 2 vertebral) and 6 costal (rib) OSA . In 10 cases, the location was not specified.

 

An increasingly considered modeling approach in cancer biology and therapeutic development is the study of naturally occurring cancers in pet dogs (referred to as comparative oncology).

The higher incidence rate of canine OSA makes the pet dog population a good model for human disease. While there is potential for any dog to develop OSA, a subset of larger dog breeds are at increased risk of developing OSA²². Differences in disease prevalence and the more aggressive disease biology in the dog further argues the opportunity for this approach to inform our understanding of this highly aggressive pediatric and non pediatric cancer³⁰.

A national canine cancer registry does not exist in the United States, European registries are extremely recent to provide accurate estimates, and most cases of suspected OSA are not definitively diagnosed histopathologically. Nevertheless, the incidence of OSA seems to be greater in dogs than in any other species, possibly 10–50 times higher than in humans⁵,²².

Although not as well characterized as in people, ample evidence exists implicating the involvement of genetic and heritable factors for the development of OSA in dogs. Currently the most thoroughly described gene mutation, which contributes to OSA formation and/or progression in dogs, is P53²³.

The commonly affected p53 tumor suppressor pathway has similar alterations in human and canine OSA²⁴. Because human TP53 is more similar to dog than mouse²⁵, and because mutations occur naturally in dogs, the canine OSA model is presumably more relevant to humans than other animal models²³.

This study included 67 dogs with OSA, and the demographic data of these dogs are similar to previous studies²⁶.

Results show that there is a significant breed effect in OSA aetiology, particularly for Rottweilers and Irish Wolfhound and, to a lesser extent, for Pyrenean Mountain Dogs, Dobermanns, Beaucerons, Belgian Shepherds, Irish Terriers, Chow Chows, Greyhounds and Leonbergers⁹,¹⁶.

Cancer is a genetic disease characterized by a progressive accumulation of genomic aberrations that are sometimes augmented by predisposing germline mutations²⁷. Although evolutionary theory has proven to be powerful for understanding cancer progression, evolutionary processes are quite different in cancers versus in speciesin ways that are important to phylogenetic inference²⁸.

When regrouping breeds by phylogenetic proximity and comparing the prevalence confidence intervals between groups, it appears that groups 4 and 5 are significantly more affected than others²⁹,³⁰. This observation suggests that the effect of genetics in OSA prevalence and in phenotype cannot be rejected.

However, only 4 out of 10 most affected breeds (Rottweiler, Irish Wolfhound, Greyhound and Dobermann) were classified into groups 4 and 5. The phylogenetic tree used in order to construct groups was built using single nucleotide polymorphisms typed in 92 breeds, leading to the definition of the groups in Table 1, and leaving some breeds as not included in any group, among which the four referred to above²⁸. This phylogenetic tree shows some differences with the group classifications of the American kennel club but also some concordances.

If we referred to the American kennel club (2018)³¹, Pyrenean Mountain Dogs and Leonbergers, which are among the most affected breeds, were not classified in our report, and could be classified into group 5. In consequence, this observation suggests that a large majority of the most affected breeds in our group studied, are phylogenetically close and classified in groups 4 and 5, but in the opinion of the authors more research is needed to confirm this hypothesis.

Our preliminary results strongly support the hypothesis that canine OSA has a major genetic aetiology. As OSA was reported in all groups, it is possible that one or many ancestral alleles involved in OSA development are old, preceding the groups differentiation, and thus transmitted to all groups of dogs, or that common environmental factors are involved. Furthermore some allele mutations increasing the OSA prevalence appeared later after the groups differentiation but before and after the split between groups 4 and 5⁸,¹⁴. Indeed, even if this report does not demonstrate differences in OSA prevalence between breeds within a group (other publications show significant differences between breeds in the group 4, particularly between Greyhound and Rottweiler, hence occurrence of new mutations at breed differentiation cannot be rejected⁹,¹⁶.

Furthermore, in group 1 the prevalence of OSA was not significantly higher than in other groups, but Belgian Shepherd, which is one of the most affected breeds, is present in this group and Beauceron, which is among the most affected breeds and was not classified in any group in this study, even though it could be classified into group 1 following the American Kennel club recommendations (2018)³¹. Moreover, groups 1 and 5 are phylogenetically closer to each other than all others groups and no significant differences of OSA prevalence was observed between groups 1, 4 and 5. Those observations support that the group 1 could potentially be predisposed to OSA even if this report does not demonstrate it.

Breed creation inadvertently selected many “founder” mutations that are associated with specific traits and diseases; this translates into reduced disease and genetic heterogeneity, consistent with the fact that most breeds are predisposed to a distinct set of diseases. Because linkage disequilibrium is up to 100-fold greater in dogs than humans, single breeds are powerful subjects for broad genetic mapping²³,³³.

Dogs are an excellent example of high genetic diversity between breeds but low genetic diversity between dogs belonging to the same breed³⁰,³⁴,³⁵. Even if the genetic diversity in humans is approximatively the same than in dogs, human populations or ethnic groups are open genetic pools, which leads to high genetic diversity between persons in a population or ethnic groups³³,³⁵. This is one of the reasons why it is difficult to find pathogenic alleles in humans.

In human medicine has been observed how OSA prevalence changes between different pools of populations, with a marked difference depending on their area of origin. Infact, OSA incidence is highest among Asian/Pacific Islanders (5.3 per million) followed by Africans and African-americans (5.1 per million), Hispanics (4.9 per million), whites (4.4 per million), and American Indian/Alaskan natives (3.0 per million)³⁶.

Despite this, due to author’s knowledges there is still no sufficient data about the relationship between ethnicity and human OSA. In support of the potential importnce of this aspect, we think that canine breed model can be higly important as study model for human OSA epidmiology.

In dogs, the genetic diversity between individuals in one breed is low and it is less difficult to identify the pathogenic alleles. Therefore, when a genetic pathology is similar in humans and dogs, it can be more interesting to start research in dogs, and particularly in the most affected breeds³²,³³.

A study comparing the transcriptional profiles of canine and human OSA found them to be virtually indistinguishable³⁰ and many of the key genes implicated in the pathogenesis or progression of OSA in children have also been characterized in the canine disease, notably PTEN: phosphatase and tensin homolog, Rb: retinoblastoma, ezrin; villin-2, c-met: mesenchymal-epithelial transition factor, erbB-2: v-erb-b2 erythroblastic leukemia viral oncogene homolog 2, neuro/glioblastoma derived oncogene homolog (avian) and  tumor protein 53³³

In 2005 the first public draft of the canine genome sequence was released³⁷,³⁸. This milestone provided the opportunity for dogs with tumors to lend additional insight into the biology of human cancers, and to more rigorously evaluate and translate novel approaches to human trials²¹.

Canine OSA resembles that of human in other several aspects. Both species species develop these tumors most commonly in the metaphysis of long bones, with micro metastases at the time of diagnosis, and overt lung metastases as the main cause of mortality³⁹,⁴⁰.

In our report, we demonstrated that 10 phylogenetically close breeds might have a higher OSA prevalence than other breeds, especially the Rottweiler and the Irish Wolfhound. A genome-wide comparison between affected dogs and non-affected dogs in those breeds could allow identification of pathogenic alleles in dogs, and potentially in humans.

Karison et al. (2013)⁴¹ compared the genome of affected dogs with the genome of a control group in Rottweilers, Irish Wolfhounds and Greyhounds. They found 33 loci implicated in heritable OSA. Moreover, loci were in majority located at regions of reduced heterozygosity, according to the hypothesis of selected pathogenic allele during the artificial selection³³,³⁵. Given the high number of loci involved in OSA prevalence, the hypothesis of a polygenic pathology could be considered⁴¹. In veterinary and human literature is possible to observe that the prevalence of OSAcan vary depending on breed, sex and age³.

As the biological behavior is similar in dogs and humans, common risk factors for developing the disease can be expected. Thus, data on incidence related to exact breeds of dogs might help define the supposed interaction between birth weight, growth rate, adult body weight or height, and the development of OSA. In this context, recognizing breeds of similar stature having significantly different incidence rates of OS is of particular interest²².

Indeed, the link between appendicular and axial OSA is higher for the bigger breeds than for the small breeds and appendicular OSA appear to be less aggressive in small breeds¹⁷. Costal (rib) OSA seems to be a different type of OSA, appearing in younger animals and, like axial OSA, preferably in female dogs¹⁹. Those differences associated with the different OSA prevalence inside groups support the existence of three different types of OSA, costal, axial and appendicular, which partially originate from mutations of different loci in the formation of groups and breeds. More precisely, in research of implicated loci, studies should be done in only one type of OSA.

Genetic similarity is the principal common point between most affected breeds, but other common factors were observed such as height¹⁴. In our report, there is a tendency to an increased OSA prevalence with increasing height but the real difference was noted between smaller (<65cm) and taller breeds (>65cm).

There is a small difference in OSA prevalence between tall breeds and it is not the tallest breeds that have the highest OSA prevalence. Some tall breeds have a lower OSA prevalence than some smaller ones. Nevertheless, in the veterinary literature, for a given breed, no correlation was reported between height and the OSA prevalence³¹. Our preliminary results associated to the literature do not allow us to considered height as a predisposing factor for OSA.

In humans, Longhi et al, suggested that hat OSA may be associated with a taller stature at least in growing individuals⁴². Despite these elements, it continuous to be hard for human epidemiological studies, to show a strong and significant correlation between body weight or height and risk of developing OSA⁴³.

Fast growing has also been supposed to be a predisposing factor for OSA³. Indeed, in large breeds the appendicular skeleton is the most affected (85%)and is subject to high remodelling during growth⁴. This parameter is highly related to the height and, as observed previously, some dogs with higher height and higher bone remodelling during growth have lower OSA prevalence than other breeds which have lower height and bone remodelling during growth¹⁶.

Moreover, in small breeds, there is 66% of axial bone OSA even if the appendicular skeleton is subject to higher remodelling than axial skeleton during growth¹. The effect of those phenotypic and physiologic parameters is ambiguous and not demonstrated^11,12. According to our results, those parameters are potentially correlated with phylogenetic proximity rather than predisposing factors. It is also possible that the oncogenic alleles are located close to alleles involved in the high height and/or high weight and/or fast growing and are therefore preferentially transmitted at the same time⁴¹. This hypothesis is supported by the report of Karisson et al. (2013)⁴¹ which have found some loci involved in heritable OSA close to loci known to control bone remodelling. It could explain the fact that the most affected breeds were large breed dogs with no correlation between those parameters and OSA prevalence when comparing dogs for a given breed or when comparing the most affected breeds (taller) between them⁵,⁷.

OSA is an infrequent tumor¹¹; hence, few cases per breed have been listed and other biases could make the prevalence lower or higher for some breeds. For example, the majority of affected dogs are geriatric, between 7 and 13 years, but dogs in the control group cover a larger span of ages, which means that some dogs from the control group might become affected later, leading to an underestimation of the observed prevalence per breed (late onset disease)³,⁷.

Due to this aspect, a smaller peak in age incidence at 18 to 24 months corresponds with the human peak incidence at late puberty, which has led to the hypothesis of skeletal growth parameters representing some of the possible etiological factors for developing this disease⁴³,⁴⁴.

Moreover this report is realised within a given geographical location where a significant difference in breed proportions could be observed in comparison with the overall population; thus it could be not representative for some breeds at a larger scale than the local scale, and should therefore be only considered as indicative. In consequence, differences between observed OSA prevalence should be considered with care due to the potentially biased nature of the sampling process⁹,¹⁵. Prevalence confidence intervals which takes into account the total number of dogs per breed, can lead to a better inter-breed comparison, although the biased sampling concerns remain.

These epidemiological aspects could be useful for establish as well as in veterinary medicine a parallelism in human medicine between OSA prevalence. In fact canine model and breeds epidemiology can represent the starting point for numerous investigations about prevalence of this bone cancer in different human populations heterogeneous for origin, ethnicity, geographical location ect.

Furthermore dogs and humans cohabit in the same environment and the push to select/hyper select some breeds could have consequently select resistant genes.

 

This preliminary report supports that OSA can be induced by one or many ancestral alleles segregating between breeds in different proportions, and is modulated by mutations occurring during the formation of groups and/or breeds and by external factors which are still misunderstood such as neutering, fast growing, intensive sport and bone fracture repair. We observed that the most affected breeds are phylogenetically closer than others and we highlighted the potential importance of genetics as an OSA aetiology. OSA in dogs shows striking similarities in tumour biology. Therefore, canine patients offer a unique opportunity for autochthonous tumour studies.

Further studies should be carried on in order to confirm these findings and to define potential oncogenic alleles and to confirm that research on canine breeds can be an important model of humane OSA research.

 

All data generated or analysed during this study are included in this article

 

Not applicable

 

Not applicable

 

The authors declare that there were no funding and support.