CT-Derived Hounsfield Units Correlate with Dorr Classification and Objectively Identify Dorr Type C Morphology in MAKO-Assisted Total Hip Arthroplasty

Every year, millions of people around the world replace their worn-out hip joints with artificial ones. Before a surgeon can choose the right metal stem to fit inside the thigh bone, they must understand the quality of that bone. If the bone is soft and the canal is wide, a standard press-fit implant might slip or break the bone during surgery. To judge this, doctors have long relied on a system called the Dorr classification. By looking at standard X-rays, they categorize the hip into three types: A, B, or C. Type A bones are dense and narrow, offering a tight fit for the new stem. Type C bones are fragile, with thin walls and a wide canal, posing a higher risk for complications. However, looking at a flat X-ray is a matter of human judgment. Two doctors might look at the same image and disagree on whether a bone is Type B or Type C. This subjectivity can make preoperative planning uncertain, especially when the difference between a safe fit and a risky one is subtle.

A team of researchers at the University of Missouri-Kansas City and St. Luke's Hospital set out to make this decision more objective. They turned to a technology that is already part of the modern surgical workflow for many patients: the preoperative CT scan. While these scans are often used to guide robotic arms during surgery, the images contain a hidden layer of data that doctors rarely use for planning. Inside every CT image, each tiny spot of bone has a number assigned to it, known as a Hounsfield unit. This number measures how much the bone blocks X-rays, which directly correlates to how dense and strong the bone is. The researchers asked a simple question: could these numbers, which are already sitting in the computer, tell them exactly which Dorr type a patient has, removing the guesswork from the equation?

To find the answer, the team looked back at the records of 489 patients who had undergone robotic-assisted hip replacement surgery between 2019 and 2024. For each patient, they had a preoperative CT scan and a standard X-ray that had already been used to assign a Dorr classification. The researchers used software to measure the density of the bone in three specific areas: the fourth lumbar vertebra in the lower back, the neck of the thigh bone, and the region between the two bumps at the top of the thigh bone. They calculated an average density score for each patient and compared it to the Dorr type assigned by the surgeons.

The results showed a clear, step-by-step relationship between the numbers and the bone types. As the bone quality worsened from Type A to Type C, the density numbers dropped steadily. Patients with the strongest, Type A bones had an average density score of 186.5. Those with Type B bones scored 136.6, and patients with the weakest, Type C bones scored just 92.2. The difference was so distinct that the researchers could use these numbers to predict the bone type with high accuracy. When they looked at the average score across all three measurement sites, they found a specific cutoff point: any patient with a score of 117.6 or lower was highly likely to have the risky Type C bone morphology. This single number correctly identified 82.2 percent of the Type C cases while correctly ruling out 80.2 percent of the safer Type A and B cases.

The study also revealed that looking at just one spot, the femoral neck, was not as effective as looking at the average of all three sites. While the femoral neck alone could identify some Type C bones, it missed many others. By combining the measurements from the spine and two parts of the hip, the researchers created a more complete picture of the patient's skeletal health. This composite score remained a strong predictor even when the researchers accounted for factors like age, sex, and body weight. In other words, the bone density numbers provided information that age or gender alone could not. The data suggested that a patient might be older or female, which are known risk factors for weaker bones, but the actual CT numbers told the true story of that specific individual's bone quality.

The researchers also examined what happened after the surgery. In this group of nearly 500 patients, six experienced a fracture of the bone around the new implant. Every single one of these fractures occurred in patients who had both the Type C bone shape on their X-rays and low density numbers on their CT scans. No fractures occurred in patients who had low numbers but a different bone shape, or in those with a Type C shape but higher density numbers. While the number of fractures was too small to draw definitive statistical conclusions about cause and effect, the pattern was striking. It suggested that combining the visual shape of the bone with the objective density numbers could help surgeons identify the very highest-risk patients before they ever entered the operating room.

This work does not propose replacing the Dorr classification or the surgeon's judgment. Instead, it offers a tool to support them. Because the CT scans are already taken for robotic planning, the density numbers are available without any extra radiation, cost, or time. The study suggests that a surgeon could look at a patient's preoperative plan and see a number like 117.6. If the number is at or below that threshold, it serves as a clear, objective flag that the bone is fragile, confirming the visual impression of a Type C hip. This could prompt the surgeon to choose a different type of implant, perhaps one that uses cement to secure the stem, or to prepare for a more delicate procedure. The goal is not to let a computer make the decision, but to give the surgeon a precise, data-driven second opinion that reduces the uncertainty of looking at a flat image. By turning a subjective visual assessment into an objective measurement, this approach aims to make the planning for hip replacement surgery more reliable and safer for patients with fragile bones.