Determining the exact location and position of a target inside the human body is one of the most critical challenges in modern medicine. Whether a surgeon is placing an implant, a radiation beam is targeting a tumor, or a radiologist is guiding a biopsy needle, the margin for error is often less than a millimeter. Getting this right depends on a layered system of imaging, tracking, physical stabilization, and coordinate mapping that work together in real time.
How the Body Gets Its Coordinates
Before any procedure can pinpoint a location inside you, clinicians need a shared language for describing where things are. Medical imaging systems use a standardized patient-based coordinate system that works like a three-dimensional map of your body. The x-axis runs left to right, the y-axis runs front to back, and the z-axis runs from your feet toward your head. Every point inside your body can be described as a set of three numbers along these axes.
This coordinate system is built into the DICOM standard, the universal format for medical images like CT and MRI scans. When your scan is loaded into surgical planning software or a radiation treatment system, every pixel in the image carries spatial information tied to this coordinate grid. That consistency is what allows a tumor identified on a preoperative MRI to be matched to the same spot during a live procedure days or weeks later.
Imaging: MRI vs. CT for Pinpointing Targets
MRI and CT scans are the two primary tools for determining where something is inside the body, and each has distinct strengths. MRI excels at showing soft tissue, making it possible to see the actual target structure directly. CT scans, by contrast, are better at showing dense objects like metal implants or bone, but the surrounding anatomy is far less visible.
A study comparing these two approaches during deep brain stimulation surgery found that the average difference in localization between intraoperative MRI and postoperative CT was less than 1 millimeter for trajectory measurements, with a mean error of 0.78 mm. Vector errors (accounting for all three dimensions) averaged 1.57 mm. These are remarkably small numbers given the complexity of the brain, but they matter: in deep brain stimulation, even a 1 mm shift can mean the difference between effective symptom relief and side effects.
MRI is generally considered closer to ground truth for defining a patient’s unique anatomy. It also has an advantage during surgery because it can account for brain shift, the slight movement of brain tissue that occurs once the skull is opened. CT scans taken immediately after surgery tend to agree closely with intraoperative MRI, but the longer the delay between the procedure and the CT scan, the greater the discrepancy between the two.
Tracking Instruments in Real Time
Knowing where a target is on a scan is only half the problem. During a procedure, clinicians also need to know exactly where their instruments are relative to that target, in real time. Two main tracking technologies make this possible: optical and electromagnetic systems.
Optical tracking uses infrared cameras that detect reflective or light-emitting markers attached to surgical instruments. By capturing these markers from multiple angles, the system calculates the instrument’s position and orientation in three-dimensional space, similar to how your eyes use two slightly different views to judge depth. The limitation is that the cameras need a clear line of sight to the markers. When markers must be placed far from the working tip of an instrument (to stay visible to the cameras), a lever-arm effect can reduce accuracy.
Electromagnetic tracking solves the line-of-sight problem by using magnetic fields instead of light. A small sensor coil on the instrument detects its position within a generated magnetic field, so it works even when the tool is hidden inside the body. This makes it particularly useful for procedures involving flexible instruments or work deep inside body cavities. The tradeoff is that nearby metal objects can distort the magnetic field and introduce errors.
Fiducial Markers as Internal Reference Points
When a target moves with breathing or shifts between imaging sessions, clinicians often place tiny metallic markers called fiducials directly into or near the target tissue. These markers are visible on imaging scans and serve as fixed reference points that tracking systems can lock onto.
Gold markers are the standard, typically just 1.0 mm wide and 3.0 to 5.0 mm long. For certain lung tumors, even thinner markers (0.4 mm wide) can be placed through specialized bronchoscope guidance. Once in place, these markers allow treatment systems to track the target’s position continuously. In robotic radiosurgery, for example, the system synchronizes the marker’s movement with the patient’s breathing pattern. External sensors (like LED markers on the chest) monitor the breathing cycle, and the system predicts where the internal marker is at any given moment, adjusting the radiation beam in near real time to follow the tumor’s path.
Sub-Millimeter Precision in Radiation Therapy
Stereotactic radiosurgery represents the most demanding application of positional accuracy in medicine. These systems deliver highly focused radiation beams to small targets, often in the brain or spine, and the clinical standard requires all mechanical components to align within less than 1 mm. The warning threshold is typically set at 0.75 mm, meaning anything above that triggers a review.
Not all machines perform equally. Testing across different linear accelerators showed significant variation. Older systems exceeded the 1 mm tolerance, averaging around 1.06 to 1.09 mm of error. Newer systems with high-definition beam-shaping components and six-degree-of-freedom positioning couches achieved accuracy of 0.41 mm at the central target point, with performance staying below 0.59 mm even at positions 3 to 7 centimeters away from center. That level of precision means the radiation beam hits within roughly the width of a few human hairs of its intended target.
Keeping the Patient Still
Even the most precise imaging and tracking systems are undermined if the patient moves. For head and neck treatments, thermoplastic masks are molded to the patient’s face and locked to the treatment table. These masks are tested at four key contact points: the forehead, both cheekbones, and the chin. Studies using daily imaging verification have measured the resulting setup errors at remarkably small values: translational shifts averaging 0.32 mm side to side, 1.09 mm front to back, and 2.24 mm head to foot, with rotational errors all under half a degree.
For targets in the chest and abdomen, breathing creates a moving target problem that immobilization alone cannot solve. Five main strategies address this. The simplest is adding a safety margin around the tumor during planning to account for its full range of motion. More active approaches include abdominal compression plates that restrict breathing depth (useful when tumor motion exceeds 5 mm), breath-hold techniques where the patient holds a deep breath during each short burst of treatment, respiratory gating that fires the beam only during a specific phase of the breathing cycle, and real-time tumor tracking that continuously adjusts the beam to follow the moving target.
Deep-inspiration breath hold is often preferred for tumors near the heart because a full inhale naturally pushes the heart away from the treatment area, reducing the dose to cardiac tissue while also temporarily freezing the tumor in a predictable position.
Where Accuracy Still Falls Short
Despite these technologies, some procedures remain challenging. Robotic needle placement for biopsies, for instance, shows a notable gap between controlled testing and real tissue. In phantom (artificial target) testing, robotic biopsy systems achieved a mean positional error of 4.34 mm. In actual tissue, that error jumped to 10.81 mm, with orientation errors nearly doubling as well. The difference comes from tissue deformation: unlike rigid phantoms, real organs compress, shift, and resist the needle unpredictably.
This gap highlights why determining exact position is not a single technology problem but a systems challenge. The coordinate framework, the imaging, the tracking, the immobilization, and the compensation for biological movement all have to work together. When one link in that chain introduces even a small error, it compounds through the rest. The trend across every subspecialty is toward tighter integration of these layers, closing the gap between where a target appears on a screen and where it actually sits inside a living, breathing body.