top of page

Explainer - Percentiles, Quartiles, T and Z Scores - What Do They Mean?

Percentiles, Quartiles, T and Z Scores - What Do They Mean?

Understanding key statistical concepts like percentiles, quartiles, T scores, and Z scores is vital in medical diagnostics and imaging for accurate test result interpretation. This article aims to clarify these terms for the general public. Medical imaging tests like CT calcium scores of the heart and DXA scans for bone density often use these metrics with a standard population to determine disease risk. Quartiles and tertiles provide alternative methods for evaluating subjects into different groupings. A clear understanding of these measures can greatly assist in interpreting various medical tests, thereby improving the precision and clarity of diagnoses.


Section I – Percentiles

Most people might be familiar with percentile scoring, such as when it's used to monitor a child's growth in terms of height, head circumference, and weight. Tracking these scores can indicate if a child is growing adequately. Additionally, these scores can show where a child stands compared to the general population of the same age and gender. Similar concepts have also been applied to medical imaging tests.


Percentiles in Calcium Scoring for CT Scans of the Heart

The percentile score is commonly used in medical imaging, particularly for calcium scoring in heart CT scans. This score aids doctors in evaluating the quantity of calcified plaque in the coronary arteries. It is calculated based on the volume of calcium in these arteries and compared to a database of scores from a population of the same age, ethnicity and gender. The score indicates the volume of your calcified plaques, which is then converted into a percentile.


“If your calcium score is in the 70th percentile, it means that 70% of individuals in the reference group have lower scores, indicating a higher risk of coronary artery disease.”

While a higher score indicates an increased risk of coronary heart disease, a low score doesn't have the same interpretation in a young patient as it does in an older one. For instance, a 45-year-old white male with a score of 10 is in the 83rd percentile, while a 65-year-old white male with the same score is in the 28th percentile.





For a deeper understanding of calcium scoring in cardiac CT scans, one can refer to the study by (Dobrolinska et al., 2021), which investigates the accuracy of risk categorization using visual and automatic AI calcium scoring methods.


Organ Volumetry

Percentiles can also be used in medical imaging to measure other body parts, such as brain volume or liver size. These measurements can be compared to a standard population to identify any abnormalities. This method is helpful for tracking changes over time and comparing an individual's measurements to a standard reference group. For example, tracking changes in brain volume can be particularly insightful.

"Although a reduction in brain volume occurs with normal ageing, an accelerated decrease in volume, especially in certain areas of the brain, could indicate pathologies like Alzheimer's disease."

Section II - Quartiles / Tertiles

In a population, the measured value (e.g. amount of visceral fat we have or the level of physical activities that we do) varies among individuals but we can usually collect samples from large cohorts to derive a representative distribution. We can then separate those into different groups. These can be in terms of tertiles (which divide the data into three equal parts) or quartiles (which divide the data into four equal parts).


Quartiles in VAT on DXA Scans

In the context of visceral adipose tissue (VAT) assessment on Dual-Energy X-ray Absorptiometry (DXA) scans, we can separate subjects into quartiles in terms of the amount of VAT they have. The first quartile represents the lowest amount of VAT and the fourth the highest. This classification aids in risk stratification and can be used for risk stratification for diseases (e.g. fatty liver or metabolic syndrome).

Studies such as those by (Spadaccini et al., 2020) and (Chan, Yu, Huang, & Vardhanabhuti, 2023) provide valuable reference values and insights into the application of quartiles in VAT assessment on DXA scans.


Tertiles – Application in Activity Tracking

Tertiles divide data into three equal parts, and in the realm of activity tracking, this can be used to differentiate individuals based on their weekly physical activity duration into 3 groups.

The World Health Organization (WHO) initially suggested 150 to 300 minutes of weekly physical activity for maintaining good health (Bull et al., 2020). Further research indicates that increased activity levels are linked to improved health outcomes, such as reduced risk of chronic diseases and all-cause mortality. Adopting this insight, Snowhill Science Limited's research teams used tertiles to categorize participants into three levels of physical activity.


The categorization method divides individuals into three groups. Those who engage in less than 168 minutes of moderate-to-vigorous physical activity (MVPA) per week are classified as 'low MVPA'. Those participating in physical activity between 168 and 356 minutes per week fall into the 'medium MVPA' category. Finally, those who engage in more than 356 minutes of MVPA per week are classified as 'high MVPA'.





Using tertiles to categorize physical activity levels gives researchers and healthcare providers a clear distinction among participants. It also enables statistical analysis for disease risk stratification. This approach offers a nuanced understanding beyond a mere binary classification. It highlights a gradient of physical activity levels, which is crucial for targeted health interventions and personalised advice.


Section III - T and Z Scores

T and Z scores are standard statistical measures used in interpreting bone density scans conducted using DXA. These scores are crucial in diagnosing osteoporosis and assessing fracture risk.


T Score in Bone Density Scans on DXA

The T score compares a patient’s bone density to the average peak bone density of a healthy 30-year-old adult. A T score of -1.0 or above is considered normal, between -1.0 and -2.5 indicates osteopenia (low bone mass), and -2.5 or below suggests osteoporosis.


Z Score in Bone Density Scans on DXA

The Z score compares a patient’s bone density to the average bone density of others their same age and sex. It is particularly useful in assessing bone density in children, premenopausal women, and men under 50.


Bazzocchi et al. (Bazzocchi et al., 2015) discuss these scores in the context of DXA scans, while Negredo et al. (Negredo et al., 2012) highlight the clinical implications of T scores in monitoring the progression to osteopenia/osteoporosis.


In conclusion, understanding these statistical measures is pivotal in the interpretation of various diagnostic tests. Percentiles provide a relative standing in a reference population, quartiles and tertiles categorize data into equal segments, and T and Z scores offer a standardized way to interpret bone density. These tools collectively enhance the precision and clarity of medical diagnostics, aiding clinicians in making informed decisions.



References:

  1. Bazzocchi, A., Ponti, F., Diano, D., Amadori, M., Albisinni, U., Battista, G., & Guglielmi, G. (2015). Trabecular bone score in healthy ageing. British Journal of Radiology, 88(1052). doi:10.1259/bjr.20140865

  2. Bull, F. C., Al-Ansari, S. S., Biddle, S., Borodulin, K., Buman, M. P., Cardon, G., . . . Willumsen, J. F. (2020). World Health Organization 2020 guidelines on physical activity and sedentary behaviour. 54(24), 1451-1462. doi:10.1136/bjsports-2020-102955 %J British Journal of Sports Medicine

  3. Chan, B., Yu, Y., Huang, F., & Vardhanabhuti, V. (2023). Towards visceral fat estimation at population scale: correlation of visceral adipose tissue assessment using three-dimensional cross-sectional imaging with BIA, DXA, and single-slice CT. 14. doi:10.3389/fendo.2023.1211696

  4. Dobrolinska, M., Lazarenko, S., Van Der Zant, F., Does, L., Prakken, N., Greuter, M., . . . Knol, R. (2021). Assessment of visual, manual, and automatic coronary calcium scoring methods - analysis of dedicated CT scans and low dose CT scans acquired during cardiac 13N-ammonia-PET/CT. European Heart Journal - Cardiovascular Imaging, 22(Supplement_3). doi:10.1093/ehjci/jeab111.017

  5. Negredo, E., Bonjoch, A., Gómez-Mateu, M., Estany, C., Puig, J., Perez-Alvarez, N., . . . Clotet, B. (2012). Time of Progression to Osteopenia/Osteoporosis in Chronically HIV-Infected Patients: Screening DXA Scan. PLOS ONE, 7(10), e46031. doi:10.1371/journal.pone.0046031

  6. Spadaccini, D., Perna, S., Peroni, G., D’Antona, G., Iannello, G., Faragli, A., . . . Rondanelli, M. (2020). DXA-Derived Visceral Adipose Tissue (VAT) in Elderly: Percentiles of Reference for Gender and Association with Metabolic Outcomes. 10(9), 163.

8 views0 comments
bottom of page