Meet The Author
March 2026
Jeremy Mortier, DVM PhD CEAV(IM) CES(CP) CPS(Ed) DipECVDI FHEA MRCVS
Hello! I am a French radiologist who loves clinical work, clinical research and teaching residents. I have mostly worked in universities (Liverpool, Paris) with short interludes in private practices. Outside work, I enjoy a quiet life with my partner, my friends and my little dog Tina. Often in a remote cabin lost in the woods, away from chaos and people ;)
AND
Diego Pulido Vega, Thomas Dalmon, Christelle Maurey, Jeremy Mortier
The Study Background
Over the course of my career in small animal radiology, I realised that my knowledge and the evidence-based medicine on the radiographic appearance of the different types of uroliths were quite limited. After discussing with specialists in internal medicine and imaging, it appeared clear that we all had various taught and/or experience-based imaging and clinical characteristics that we used to try to differentiate the types of uroliths. Considering the difference in treatments to remove uroliths based on their composition, and the invasive methods to collect them for analysis, a more robust method of differentiation would be extremely useful.
What is the primary knowledge gap your study aims to address?
Evidence-based medicine about the radiographic appearance of uroliths based on their composition is often old, performed in vitro, and scarce. Studies on the in vivo radiographic appearance of uroliths based on their mineral composition in dogs and cats were long overdue. We run two studies, one in dogs (doi: 10.1111/jvim.70252) and one in cats (https://doi.org/10.1093/jvimsj/aalaf005).
The Study Design
Both articles have a similar study design. They are retrospective cross-sectional studies
- 202 dogs with lower urinary tract uroliths composed of ≥ 70% calcium oxalate (n = 109), struvite (n = 58), or cystine (n = 35).
- 69 cats with lower urinary tract uroliths containing ≥70% of calcium oxalate (n = 45) or struvite (n = 24).
Radiographs were evaluated for urolith size, shape, surface, borders, internal architecture, and quantitative radiopacity, and also assessed using a ratio of urolith-to-L5 vertebra pixel values (UOR). The diagnostic performance of urolith size and UOR was assessed with receiver operating characteristic (ROC) curves, calculating the area under the ROC curve (AUC).
What are the main study results?
In dogs:
- Calcium oxalate uroliths were often highly radiopaque, associated with concurrent nephroliths, had unique shapes such as bosselated or spiculated, and were less commonly ovoid.
- Struvite uroliths were larger (median: 15 mm; IQR: 8.1–25.8 mm) with a cut-off of ≥ 11 mm (AUC = 0.82, p < 0.001; Se 91%, Sp 67%) and were associated with pyramidal shapes or solitary cystoliths.
- Cystine uroliths had lower radiopacity (median UOR: 0.71; IQR: 0.63–0.80), with a cut-off of UOR ≤ 0.83 (AUC = 0.81, p < 0.001; Sp 67%, Se 84%).
In cats:
- Struvite uroliths had lower opacity with a UOR cut-off ≤ 0.78 (AUC = 0.81, P < .001; Se 58%, Sp 92%), were larger, with a cut-off value of ≥ 7.3 mm (AUC = 0.75, P < .001; Se 38%, Sp 98%), more frequently irregularly shaped, with rough surface and heterogeneous architecture.
- Urease-positive bacteriuria was present in 43% of struvite uroliths ; infected struvite uroliths were significantly larger (median, 9.8mm ; interquartile range [IQR], 5.4-13 mm) than sterile ones (median, 4.6 mm; IQR, 2.6-5.4 mm; P = .02).
- Calcium oxalate uroliths were frequently smaller, more radiopaque, uniform, smooth, and associated with nephroliths.
Were there any unexpected results or challenges during your research?
By lack of oversight, we originally didn't take into account the presence or absence of urinary tract infections in the radiographic appearance of the uroliths. Which means that we had to go back to patients’ files and repeat the stats :( However, it was both very interesting and unexpected to discover that infected struvites were larger than non-infected ones in cats (no difference in dogs).
Takeaways from this study
These studies provide a number of radiographic features that, in cunjunction with the animal epidemiological and clinical data, help predict the nature of commonly encountered lower urinary tract uroliths in dogs and cats.
What future directions would you like to explore based on this study?
The next step would be to develop an algorythm or a deep learning tool that would incorporate the epidemiological, clinical and in vivo imaging data of an animal in order to accurately predict the mineral composition of lower urinary tract uroliths.
