When clinicians evaluate a patient’s health from a medical scan, they have historically relied on simplified geometric shortcuts. To measure a tumor, they might take a single linear measurement of its longest diameter. To evaluate a patient's physical reserve or muscle wasting (sarcopenia), they might measure the cross-sectional area of muscle on a single transverse slice of an abdominal CT scan.
While these one-dimensional (1D) and two-dimensional (2D) metrics are convenient, they rest on a flawed assumption: that a single line or plane can accurately represent complex, irregular three-dimensional (3D) biological structures. In reality, tumors rarely grow or shrink as perfect spheres, and human tissue distribution varies significantly across the body.
A rapidly growing research direction shifts the paradigm from these flat proxies to 3D volumetric quantification. By segmenting and integrating voxels across entire imaging volumes, this approach captures the true spatial distribution, morphology, and total mass of tumors and healthy tissues alike.
The Failure of Flat Proxies
To understand why 3D volumetry works, we must first look at why traditional 1D and 2D measurements fail.
Consider an irregular tumor undergoing chemotherapy. As it responds to treatment, it may shrink asymmetrically, cave inward, or fragment into several smaller clusters. A linear diameter measurement taken across a single axis can completely miss this fragmentation, falsely indicating that the tumor is stable when its total mass has actually plummeted.
A similar problem occurs in body composition analysis. Traditional methods often quantify muscle and fat area at a single vertebral cross-section (such as the L3 level) and extrapolate that number to represent the whole body. However, because the human torso is not a uniform cylinder, localized cross-sectional sampling introduces significant error due to regional anatomical variations and patient positioning.
By transitioning to 3D volumetric analysis, researchers are eliminating these sampling errors. Summing up voxels across entire anatomical regions yields an objective, highly precise measurement of either the total disease load or a patient's physiological reserves.
Volumetric Tumor Dynamics: Beyond Diameter
In oncology, tracking the exact volume of cancerous tissue over time provides a far more sensitive indicator of disease progression and treatment response than traditional metrics.
As demonstrated by Xiao et al. in their study on locally advanced rectal cancer, calculating the Tumor Volume Reduction Rate (TVRR) via 3D MRI allows clinicians to capture complex morphological changes and asymmetric shrinkage. This volumetric approach accurately predicts how well a tumor is responding to neoadjuvant chemotherapy before surgery, a feat that traditional 1D or 2D diameter metrics struggle to achieve because they overlook structural fragmentation.
Traditional 1D/2D (RECIST) 3D Volumetric Integration
[ Diameter ] [Voxels Summed]
| | |||
v v vvv
( Irregular Shape ) ( Irregular Shape )
Misses asymmetric changes Captures true mass & volume
This precision is equally vital when monitoring slow-growing or rare cancers. For instance, Wennmann et al. developed a biomarker called "Speed of Growth" (SOG) for smoldering multiple myeloma. By utilizing 3D segmentation to calculate the total volume of all focal lesions across a whole-body MRI, the authors showed that tracking the rate of volumetric expansion over time provides vastly superior risk stratification compared to simply counting discrete lesions. A patient might have numerous stable lesions but a low risk, whereas a single rapidly expanding mass—caught only via volumetry—signals immediate danger.
Similarly, Fuller et al. utilized 3D volumetric segmentation to calculate precise tumor doubling times in metastatic adrenocortical carcinoma, revealing distinct growth rates across different organ sites. In advanced prostate cancer, Weber demonstrated that aggregating the total tumor volume (PSMA-TV) from 3D PSMA PET-CT scans serves as a powerful independent predictor of overall survival for patients undergoing radioligand therapy, outperforming traditional blood-based biomarkers like Prostate-Specific Antigen (PSA).
Even in early-stage cancers, volume-derived density metrics provide deeper insights. Zhao et al. showed that 3D-CT volumetric reconstructions—specifically evaluating the solid-to-tumor volume ratio and average CT density values—predict malignancy in lung adenocarcinomas far better than traditional 2D size metrics, giving surgeons an objective framework to choose between aggressive and lung-sparing surgeries.
Mapping Physiological Reserve Through Body Composition
The exact same mathematical principle—integrating voxels across a 3D space to find true mass—is revolutionizing how we evaluate a patient's baseline health, or "physiological reserve."
Instead of relying on Body Mass Index (BMI), which cannot differentiate between muscle and fat, or single-slice 2D areas, modern pipelines integrate tissue volumes across entire anatomical regions. As Lee et al. demonstrated, calculating a "Torso Muscle Volumetric Index" (TorsoMVI) by segmenting the entire torso from low-dose CT scans drastically reduces the sampling errors inherent in single-slice abdominal measurements. This holistic approach yields a much stronger correlation with actual muscle strength and clinical reference standards for sarcopenia.
This comprehensive spatial mapping has profound prognostic value across a variety of acute and chronic conditions:
- Immunotherapy Outcomes: Decazes et al. showed that 3D-measured subcutaneous fat and muscle volumes act as powerful, independent predictors of survival for lung cancer patients undergoing immunotherapy, successfully identifying high-risk groups that BMI completely misclassifies.
- Infectious Disease Severity: In the context of COVID-19, Hosch et al. developed an automated pipeline to quantify tissue volumes across the entire thoracic cavity. The authors found that these 3D sarcopenia and cardiac markers strongly correlate with disease severity and mortality.
- Pan-Cancer Stratification: Borys et al. demonstrated that a 3D sarcopenia index extracted from routine clinical imaging provides superior prognostic stratification across multiple cancer types by capturing muscle and bone distribution across entire anatomical zones.
- Morphological Health Profiling: Going beyond raw volume, Huang et al. combined 3D volumetric measurements with morphological analysis (such as the surface-to-volume ratio and elongation of muscle groups) to uncover hidden patterns in tissue distribution that predict severe complications like infectious pancreatic necrosis.
The Automation Breakthrough: Making Volumetry Practical
If 3D volumetric analysis is so clearly superior to 1D and 2D shortcuts, why hasn't it been the clinical standard for decades?
The bottleneck has always been human labor. Manually tracing the boundaries of a tumor or separating muscle from fat across dozens or hundreds of sequential radiation slices is incredibly tedious, taking a radiologist upwards of 30 to 45 minutes per scan. In a fast-paced clinical environment, this made volumetry completely impractical.
The catalyst that changed everything is deep learning. By training convolutional neural networks (such as nnU-Net) and self-learning architectures on large, expert-annotated datasets, researchers have automated the segmentation process.
For example, Bereska et al. developed COALA, a self-learning AI model designed to segment colorectal cancer liver metastases. By utilizing a "teacher-student" training paradigm, the model achieves near-human accuracy in a fraction of the time. This rapid, automated extraction of Total Tumor Volume (TTV) finally bridges the gap between research and the clinic, transforming 3D volumetry from an academic luxury into a scalable, real-time diagnostic tool.
Key Foundational Work
Volumetry based biomarker speed of growth: Quantifying the change of total tumor volume in whole-body magnetic resonance imaging over time improves risk stratification of smoldering multiple myeloma patients (2018) — contributed 2 concept elements to this direction.
Tumor Volume Reduction Rate Predicts Pathologic Tumor Response of Locally Advanced Rectal Cancer Treated with Neoadjuvant Chemotherapy alone: Results from a Prospective Trial (2015) — contributed 2 concept elements to this direction.
Radiologic Predictors for Clinical Stage IA Lung Adenocarcinoma with Ground Glass Components: A Multi-Center Study of Long-Term Outcomes (2015) — contributed 1 concept elements to this direction.