Diagnostic Value of Condylar Distance Asymmetry Relative to the Femoral and Tibial Anatomic Axes in Anterior Cruciate Ligament Rupture

Abstract

Background:

Anatomical risk factors for anterior cruciate ligament (ACL) injury remain incompletely defined, particularly those reflecting combined tibiofemoral geometry.

Purpose:

To determine whether the tibial axis–condylar distance difference (T-ACD) and femoral ACD (F-ACD) are independently associated with ACL rupture and to compare their performance with other morphometric indices.

Study Design:

Cross-sectional study; Level of evidence, 3.

Methods:

In this single-center study, 63 magnetic resonance imaging (MRI)–confirmed ACL-ruptured knees were age- and sex-matched to 62 control knees from 2022 to 2024. Weightbearing anteroposterior radiographs provided T-ACD and F-ACD; true lateral radiographs and MRI yielded posterior tibial slope (PTS), lateral tibial height (LTH), medial tibial plateau depth (MTD), lateral femoral condylar index (LFCI), LFC height-to-depth ratio (LFC-H/D) and notch width index (NWI). Two blinded observers performed all measurements and repeated them after 3 weeks; intra- and interobserver reliability were calculated. Multivariate logistic regression with intraclass correlation coefficient (ICC) and receiver operating characteristic analysis assessed independent associations and discriminative ability.

Results:

ACL rupture knees were younger than controls (27.3 ± 7.9 vs 30.3 ± 7.6 years; P = .04); sex, side, and body mass index were comparable (P > .05). Intraobserver ICC was 0.86 to 0.94 and interobserver ICC measured 0.83 to 0.92. T-ACD, F-ACD, LTH, and LFCI were significantly greater in ruptured knees (P < .001; Cohen d≥ 0.67). A T-ACD ≥4.5 mm conferred an odds ratio of 1.84/mm (95% CI, 1.34-2.53) and showed good discrimination (area under the curve [AUC] = 0.858). The composite model (T-ACD + LTH + LFCI) achieved excellent accuracy (optimism-corrected AUC = 0.957; Nagelkerke R² = 0.83). PTS, LFC-H/D, MTD, and NWI were not significant predictors (P > .05).

Conclusion:

Our study showed that coronal-plane condylar height asymmetry, particularly T-ACD, along with LTH and LFCI, may serve as potentially useful morphometric markers for ACL rupture risk. If confirmed by external validation, identification of high-risk morphometry on routine radiographs and MRI may help guide targeted neuromuscular training and injury prevention programs in at-risk individuals

Keywords

anterior cruciate ligament knee joint femur tibia rupture

Anterior cruciate ligament (ACL) injuries are among the most common intra-articular knee ligament injuries in athletes and active individuals, leading to considerable pain, instability, and a decline in quality of life.³ Identifying risk factors for ACL rupture is therefore crucial for early diagnosis and prevention. Over the past decades, numerous studies have explored these risk factors to improve our understanding, diagnosis, and treatment of ACL injuries.³

In addition to traumatic mechanisms, growing evidence indicates that knee joint morphology contributes to the etiology of ACL ruptures.⁴ Morphometric features of the tibia and femur have attracted particular interest.^4,11 Frequently cited morphological risk factors include an increased posterior tibial slope (PTS), reduced tibial eminence size, and narrowed intercondylar notch, with most research focusing on the proximal tibia.^4,11 Recently, several studies have shown that distal femoral characteristics are also significantly associated with ACL tears.⁹ In particular, femoral condylar asymmetry (the proportional relationship between the lateral and medial condyles) may increase ACL stress through condylotibial mismatch.^7,17 However, data on this topic remain limited. Likewise, parameters such as lateral tibial plateau height, PTS, and tibial condylar differences may influence ACL biomechanics, yet the literature on these factors is still sparse, especially studies employing multivariate analyses that simultaneously assess tibiofemoral morphology.

Our primary aim was to evaluate whether the tibial axis–condylar distance difference (T-ACD) and the femoral ACD (F-ACD) are independently associated with ACL rupture. Secondarily, we investigated the relationship between ACL injury and a panel of additional morphometric variables—lateral tibial height (LTH), lateral femoral condylar index (LFCI), PTS, medial tibial plateau depth (MTD), LFC height-to-depth ratio (LFC-H/D), and notch width index (NWI). We hypothesized that both T-ACD and F-ACD would be significantly greater in ACL-deficient knees than in matched controls and would remain robust predictors of rupture after adjusting for age, sex, and the aforementioned covariates.

Methods

This study was designed as a retrospective, single-center case-control imaging analysis. Institutional review board approval was obtained before data collection, and the investigation was carried out in accordance with the principles of the Declaration of Helsinki.

All knee trauma patients who presented to our hospital between 2022 and 2024 were screened. Knee radiographs and magnetic resonance imaging (MRI) were reviewed to identify cases with and without ACL rupture. Of 412 MRI examinations assessed, 287 were excluded according to predefined criteria (with the number of exclusions listed alongside each criterion); the remaining 125 examinations constituted the study cohort (Figure 1). A total of 63 patients with MRI-confirmed ACL rupture formed the rupture group, whereas 62 age- and sex-matched patients with an intact ACL and no other MRI pathology served as the control group.

Figure 1.

Study flow diagram. ACL, anterior cruciate ligament; KL, Kellgren-Lawrence grade; MRI, magnetic resonance imaging.

Demographic variables—including age, sex, affected side, and body mass index (BMI)—were recorded. BMI was categorized according to the World Health Organization classification: underweight (<18.5 kg/m²), normal weight (18.5-24.9 kg/m²), overweight (25.0-29.9 kg/m²), and obese (≥30.0 kg/m²). All morphometric measurements were subsequently performed on the imaging data sets.

Exclusion Criteria

The following exclusion criteria were applied: prior knee surgery (n = 37), multiligamentous injury (n = 15), osteoarthritis with Kellgren-Lawrence grade ≥3 (n = 32), inadequate imaging quality (n = 82), systemic connective tissue disorder (n = 4), skeletal immaturity (≤18 years) (n = 24), periarticular fracture or residual deformity (n = 23), marked coronal malalignment (n = 8), and partial or suspected ACL tears—only complete ruptures or fully intact ACLs were accepted (n = 62).

Morphometric Measurements

F-ACD: On an anteroposterior (AP) knee radiograph, the femoral anatomic axis was constructed by identifying 2 middiaphyseal reference points in the distal femur, located approximately 10 cm and 15 cm proximal to the knee joint line, respectively. A line connecting these 2 points defined the femoral anatomic axis. Next, a perpendicular reference line was drawn from the femoral axis at the level of the distal femur (Figure 2A). From this reference line, 2 perpendicular distances were measured to the most distal points of the medial and LFCs (h1 and h2). The absolute difference between these 2 distances (|h1−h2|) represented the F-ACD (Figure 2A). This metric quantifies coronal plane distal femoral condylar asymmetry relative to the femoral shaft axis.

T-ACD. Analogously, on the same AP radiograph, the tibial anatomic axis was defined by connecting 2 middiaphyseal points located approximately 8 cm and 13 cm distal to the knee joint line. A line connecting these points defined the tibial anatomic axis. Next, a perpendicular reference line was drawn from the tibial axis at the level of the tibial plateaus (Figure 2B). From this reference line, 2 perpendicular distances were measured to the most inferior points of the medial and lateral tibial condyles (h1 and h2). The absolute difference between these distances (|h1−h2|) was defined as the T-ACD (Figure 2B). This parameter captures coronal plane asymmetry in the distal levels of the medial and lateral tibial plateaus relative to the tibial shaft axis.

PTS. On the true lateral knee radiograph, the slope of the medial tibial plateau was calculated with respect to the proximal tibial anatomic axis. The PTS was defined as the angle between a line connecting the anterior and posterior edges of the medial tibial plateau and a line perpendicular to the tibial shaft axis (Figure 2C).

LFCI. Measurements were performed on the sagittal MRI slice that best visualized the LFC, with slice selection confirmed using the axial reference image (Figure 3). The anterior and posterior subchondral contours of the LFC were assumed to be approximately circular. Two best-fitting circles were constructed on the LFC: an extension circle (A) fitted to the contour between the most inferior and most anterior points, and a flexion circle (B) fitted to the contour between the most inferior and most posterior points (Figure 3). The diameter of B was divided by the diameter of the A to calculate the LFCI (LFCI = B/A).

LFC-H/D. Measurements were performed on the sagittal MRI slice that best visualized the LFC, with slice selection confirmed using the axial reference image (Figure 4). Three reference lines were drawn as shown in Figure 4: line A, a vertical line aligned with the femoral shaft axis; line C, a horizontal line tangential to the most posterior subchondral point of the LFC; and line B, a horizontal line tangential to the most distal point of the LFC. The depth (D) was defined as the distance from line A to the posterior tangent line (line C). The height (H) was defined as the distance between the distal tangent line (line B) and the posterior tangent line (line C). The ratio H/D was recorded as the LFC-H/D (Figure 4).

NWI. The width of the femoral intercondylar notch at its narrowest point was measured on the coronal MRI slice that best displayed both femoral condyles. This width was divided by the bicondylar width (the maximal mediolateral distance between the lateral and medial femoral condyles) on the same slice (Figure 5A).

MTD. In the sagittal MRI plane that optimally visualized the medial tibial plateau, a reference line connecting the anterior and posterior margins of the plateau was drawn. The perpendicular distance from this line to the deepest (most inferior) point of the plateau represented the MTD (Figure 5B). A greater depth indicates a more concave medial plateau.

LTH. The vertical distance between the articular surface of the lateral tibial plateau and the floor of the intercondylar notch (tibial eminence base) was measured in the sagittal MRI plane that best displayed the lateral plateau and intercondylar region (Figure 5C). A greater LTH indicates an elevated lateral plateau rim.

Figure 2.

Radiographic morphometric measurements. (A) Femoral axis–condylar distance difference: black line indicates the femoral anatomic axis; yellow line, the perpendicular reference line at distal femoral level; red line, the perpendicular distance to the medial femoral condyle (h1); and green line, the perpendicular distance to the lateral femoral condyle (h2). (B) Tibial axis–condylar distance difference: black line indicates the tibial anatomic axis; yellow line, the perpendicular reference line at tibial plateau level; blue line, the perpendicular distance to the medial tibial condyle (h1); and orange line, the perpendicular distance to the lateral tibial condyle (h2). (C) Posterior tibial slope (PTS): black line indicates the tibial shaft axis; green line, the medial tibial plateau surface; yellow line is perpendicular to the tibial shaft axis; and angle (red α) indicates the PTS.

Figure 3.

Magnetic resonance imaging (MRI)–based morphometric measurement of the lateral femoral condylar index (LFCI). (A) Sagittal proton density (PD) MRI slice demonstrating the LFC at its maximal anteroposterior diameter. Two best-fitting red circles are overlaid on the subchondral contour of the LFC: the larger, anterior circle (A) represents the extension circle fitted to the contour between the most inferior and most anterior points; the smaller, posterior circle (B) represents the flexion circle fitted to the contour between the most inferior and most posterior points. Yellow lines indicate the diameters of each circle. LFCI is calculated as the diameter of circle B divided by the diameter of circle A. (B) Axial PD MRI reference image used to confirm and standardize sagittal slice selection at the level of the LFC, ensuring the chosen sagittal slice represents the true maximal anteroposterior diameter of the condyle.

Figure 4.

Magnetic resonance imaging (MRI)–based morphometric measurement of the lateral femoral condylar height-to-depth ratio (LFC-H/D). (A) Sagittal proton density (PD) MRI slice demonstrating the LFC at its maximal anteroposterior diameter, confirmed using the axial reference image. Three blue reference lines are drawn: vertical line (A) aligned with the femoral shaft axis; horizontal line (B) tangential to the most posterior subchondral point of the LFC; horizontal line (C) tangential to the most distal point of the LFC. Red horizontal arrow (D) represents the depth, defined as the horizontal distance from line A to the posterior tangent line B. Red vertical arrow (H) represents the height, defined as the vertical distance between lines B and C. The LFC-H/D ratio is calculated as H divided by D. (B) Axial PD MRI reference image used to confirm and standardize sagittal slice selection at the level of the LFC.

Figure 5.

Magnetic resonance imaging (MRI)–based morphometric measurements. (A) Coronal T2 MRI slice at the level of the posterior cruciate ligament attachment demonstrating the notch width index (NWI). The red line (A) represents the width of the femoral intercondylar notch at its narrowest point; the yellow line (B) represents the maximal mediolateral bicondylar width. NWI is calculated as A divided by B. (B) Sagittal proton density MRI slice optimally visualizing the medial tibial plateau, demonstrating the medial tibial plateau depth. The red line represents the reference line connecting the anterior and posterior margins of the medial tibial plateau; the yellow line (d) indicates the perpendicular distance from this reference line to the deepest (most inferior) point of the plateau. A greater depth indicates a more concave medial plateau. (C) Sagittal T2 MRI slice optimally visualizing the lateral tibial plateau and intercondylar region, demonstrating the lateral tibial height (LTH). The white line represents the vertical distance between the articular surface of the lateral tibial plateau and the floor of the intercondylar notch; the red line (h) indicates this measured distance. A greater LTH indicates an elevated lateral plateau rim.

All cases were first identified by MRI performed to evaluate ACL integrity in patients with suspected rupture. Morphometric analysis was then carried out on 3 separate imaging sets chosen to provide the most reliable plane for each parameter: (1) weightbearing AP knee radiographs, (2) true lateral knee radiographs, and (3) knee MRI slices. All measurements were obtained with digitally calibrated picture archiving and communication system (PACS) software (Extreme PACS); linear distances were recorded to the nearest 0.1 mm and angles to the nearest 0.1°. For proportional indices (LFCI, LFC-H/D, NWI), pixel-based measurements were automatically scale corrected, so magnification effects were considered negligible.

Radiologic Standardization

Radiographic Standardization

For the weightbearing AP view, patients stood with the knees in near-full extension and the patellae facing forward. The tube-detector distance was set at 100 cm, and the central beam was directed ~1 cm distal to the joint line. Lower-limb rotation was kept to a minimum. In true lateral view, the knee was flexed to approximately 30°, and the femoral condyles were superimposed. A posterior condylar overlap ≤3 mm was accepted as a true lateral projection.

MRI Standardization

All MRI examinations were performed on a 1.5-T scanner using a standard knee protocol (sagittal PD fat-saturated, coronal T2–turbo spin-echo and axial PD sequences; 3-mm slice thickness). Slices with artifact, poor meniscal visualization, or segmentation difficulty were excluded from morphometric analysis. For each parameter, slice selection was standardized using the following anatomic landmarks: (1) LFCI and LFC-H/D: sagittal slice where the LFC demonstrated maximal AP diameter with clear visualization of both anterior and posterior subchondral margins, confirmed using axial reference images; (2) LTH and MTD: sagittal slice where the tibial plateau and intercondylar eminence were simultaneously visible with maximal clarity; (3) NWI: coronal slice at the level of the posterior cruciate ligament attachment where both femoral condyles and the intercondylar notch were optimally visualized in a single plane.

Measurement Protocol and Reliability Analysis

Using radiographic analysis in the weightbearing AP view, F-ACD and T-ACD were measured; in the true lateral view, PTS was obtained. Using MRI analysis, LFCI, LFC-H/D, MTD, and LTH were measured with sagittal slices, and NWI was measured with coronal slices.

Two independent observers (E.İ. and M.A.K.) performed all measurements on the full study cohort and repeated them after a 3-week interval to assess both intra- and interobserver reliability. Before data collection, they completed a joint training session on 20 pilot cases (excluded from the final analysis) to standardize anatomical landmarks and measurement techniques. Inter- and intraobserver reliability were quantified using a 2-way random-effects, absolute-agreement intraclass correlation coefficient, ICC(2,1), and interpreted as follows: ICC <0.50 poor, 0.50 to <0.75 moderate, 0.75 to 0.90 good, and >0.90 excellent; 95% CIs were calculated for all ICC estimates. It should be noted that ICC values quantify measurement reproducibility (consistency across repeated measurements) rather than accuracy relative to true 3-dimensional anatomy; the latter would require validation against volumetric imaging or anatomical specimens, which was beyond the scope of this study.

Statistical Analysis

All analyses were carried out with IBM SPSS Statistics Version 25.0.1 and R Version 4.3.0 (R Foundation for Statistical Computing) using the rms and pROC packages. An a priori power calculation with GPower Version 3.1 indicated that, assuming an expected effect size of d = 1.0 for the T-ACD, α =.05, and 80% power, a minimum of 25 patients per group would be sufficient; the available sample (63 ACL-ruptured knees, 62 controls) therefore provided >99% power. Continuous variables were checked for normality with the Shapiro-Wilk test and screened for outliers using a ±3 z-score threshold. As no demographic or morphometric data were missing, imputation was unnecessary. Normally distributed data are presented as mean ± SD, and nonnormal data as median (25th-75th percentile). Group comparisons employed the independent-samples t test or the Mann-Whitney U test, with effect sizes reported as Cohen d or r, respectively. Categorical variables were analyzed with the χ² test or Fisher exact test where expected counts were <5, and effect size was expressed by Cramér V. To control the family-wise type 1 error arising from 8 principal morphometric parameters, a Bonferroni correction set the significance threshold at P < .006. Potential risk factors for ACL rupture were first screened by univariate logistic regression; variables with P < .10, together with age and sex, were entered into a multivariate model using the forward likelihood ratio method. Model stability was ensured by keeping the events-per-variable ratio >10 and checking multicollinearity with a variance inflation factor <5. To assess model performance beyond discrimination, we performed internal validation via bootstrapping (1000 resamples) to obtain optimism-corrected estimates of the area under the receiver operating characteristic (ROC) curve (AUC). Model calibration was evaluated using calibration-in-the-large, calibration slope, and the Brier score. Calibration plots were constructed by grouping predicted probabilities and plotting mean predicted probability against observed proportion of ACL ruptures. Diagnostic performance was evaluated at the optimal probability threshold determined by Youden index, with calculation of sensitivity, specificity, positive and negative predictive values (PPV, NPV), positive and negative likelihood ratios (LR+, LR–), and confusion matrices. Goodness-of-fit was assessed with the Hosmer-Lemeshow test, explanatory power with Nagelkerke R², and overall accuracy with the percentage of correctly classified cases. For each continuous morphometric measure, ROC curves were generated; AUC values were interpreted as poor (0.50 to <0.60), fair (0.60 to <0.70), moderate (0.70 to <0.80), good (0.80 to <0.90), or excellent (≥0.90) according to Hosmer and Lemeshow classification.¹⁴ All tests were 2-tailed, and statistical significance was accepted at P < .05 unless otherwise adjusted.

Results

A total of 125 knee MRIs were analyzed, comprising 63 knees with ACL rupture and 62 control knees. In the overall cohort, the mean age was 28.8 ± 7.8 years and the mean BMI was 22.2 ± 2.2 kg/m². Among male patients, 47 patients (74.6 %) were in the rupture group and 37 (59.7 %) were in the control group, a difference that did not reach statistical significance (P = .08). Side distribution was similar between groups (right knee: 50.8 % vs 48.4 %; P = .79). Patients with ACL rupture were significantly younger than controls (27.33 ± 7.91 years vs 30.27 ± 7.59 years; P = .04), whereas BMI did not differ between groups (22.05 ± 2.41 kg/m² vs 22.29 ± 2.08 kg/m²; P = .55). When BMI categories were considered—underweight (<18.5 kg/m²), normal (18.5-24.9 kg/m²), and overweight (25.0-29.9 kg/m²)—their distribution was comparable (P = .62). No patient met the criterion for obesity (≥30 kg/m²) during the study period. Thus, younger age was the only demographic variable significantly associated with ACL rupture (Table 1).

Table 1

Baseline Demographic Characteristics and BMI Categories ^a

Variable	ACL Rupture Group(n = 63)	Control Group(n = 62)	P
Age, y	27.33 ± 7.91	30.27 ± 7.59	.04
BMI, kg/m²	22.05 ± 2.41	22.29 ± 2.08	.55
Sex, male	47 (74.6)	37 (59.7)	.08
Side, right knee	32 (50.8)	30 (48.4)	.79
BMI category			.62
Underweight, <18.5	7 (11.1)	4 (6.5)
Normal, 18.5-24.9	48 (76.2)	51 (82.3)
Overweight, 25.0-29.9	8 (12.7)	7 (11.3)

Data are presented as mean ± SD or n (%). Independent-samples t tests were applied to age and BMI, whereas chi-square tests (Fisher exact test when expected cell counts were <5) were used for sex, side, and BMI category distributions. All tests were 2-tailed, and P < .05 was considered statistically significant. ACL, anterior cruciate ligament; BMI, body mass index.

Morphometric analysis revealed that F-ACD, T-ACD, LTH, and LFCI were all significantly greater in the ACL-rupture group than in controls (P < .001 for each), with large clinical effect sizes (Cohen d≥ 0.67). MTD showed nominal significance (P = .02) but did not survive the Bonferroni-adjusted threshold (α = .006) and was therefore regarded as a secondary finding. PTS, the LFC-H/D, and the NWI did not differ between groups (P > .05) and were not entered into the multivariate regression model (Table 2).

Table 2

Morphometric Parameters ^a

Measurement	ACL Rupture Group(n = 63)	Control Group(n = 62)	P
Tibial axis–condylar distance difference, mm	7.11 ± 1.96	4.41 ± 1.65	<.001
Femoral axis–condylar distance difference, mm	8.92 ± 2.18	7.47 ± 2.09	<.001
Lateral tibial height, mm	4.87 ± 1.70	2.76 ± 0.88	<.001
Lateral femoral condylar index	0.715 ± 0.056	0.660 ± 0.049	<.001
Medial tibial plateau depth, mm	2.97 ± 1.08	2.58 ± 0.67	.02
Posterior tibial slope, deg	10.89 ± 3.15	10.66 ± 2.95	.68
Lateral femoral condylar height-to-depth ratio	0.486 ± 0.056	0.493 ± 0.039	.36
Notch width index	0.259 ± 0.022	0.260 ± 0.020	.77

Data are presented as mean ± SD. Independent-samples t tests were used for group comparisons; normality was verified with the Shapiro-Wilk test and homogeneity of variances with the Levene test. Variables that violated parametric assumptions were cross-checked with the Mann-Whitney U test; because significance did not change, parametric P values are reported. A Bonferroni correction was applied for multiple comparisons. ACL, anterior cruciate ligament.

Measurement Reliability

Individual ROC analyses for the significant morphometric parameters demonstrated good discriminative ability: T-ACD (AUC = 0.858), LTH (AUC = 0.854), and LFCI (AUC = 0.781; optimal cutoff = 0.698, sensitivity = 69.8%, specificity = 75.0%). MTD demonstrated poor discrimination (AUC = 0.605). Inter- and intraobserver reliability results for all parameters are summarized in Table 3.

Table 3

Inter- and Intraobserver Reliability ^a

Parameter	Interobserver	Interpretation	Intraobserver	Interpretation
F-ACD	0.92 (0.88-0.95)	Excellent	0.94 (0.91-0.96)	Excellent
T-ACD	0.91 (0.87-0.94)	Excellent	0.93 (0.90-0.96)	Excellent
LTH	0.89 (0.84-0.93)	Good	0.91 (0.87-0.94)	Excellent
LFCI	0.86 (0.81-0.91)	Good	0.88 (0.84-0.92)	Good
MTD	0.84 (0.78-0.89)	Good	0.87 (0.82-0.91)	Good
PTS	0.88 (0.83-0.92)	Good	0.90 (0.86-0.94)	Excellent
LFC-H/D	0.85 (0.79-0.90)	Good	0.87 (0.82-0.91)	Good
NWI	0.83 (0.77-0.88)	Good	0.86 (0.81-0.90)	Good

ICC values interpreted as <0.50 poor, 0.50 to <0.75 moderate, 0.75 to 0.90 good, >0.90 excellent. F-ACD, femoral axis–condylar distance difference; ICC, intraclass correlation coefficient; LFC-H/D, lateral femoral condylar height-to-depth ratio; LFCI, lateral femoral condylar index; LTH, lateral tibial height; MTD, medial tibial plateau depth; NWI, notch-width index; PTS, posterior tibial slope; T-ACD, tibial axis–condylar distance difference.

Model Validation and Performance

Internal validation via bootstrapping (1000 resamples) demonstrated robust performance with minimal optimism (Table 4). The optimism-corrected AUC was 0.957, while the apparent AUC was 0.962 (95% CI, 0.931-0.992), yielding an optimism of only 0.005, indicating excellent model stability. Calibration metrics were ideal: calibration-in-the-large = 0.000, calibration slope = 1.000 (95% CI, 0.68-1.45), and Brier score = 0.069. The calibration plot (Figure 6) confirmed close alignment between predicted and observed risks. The ROC curve (Figure 7) demonstrated excellent discriminative ability. At the optimal cutoff (0.471, determined by Youden index), diagnostic performance was excellent (Table 5): sensitivity, 90.5% (95% CI, 80.7%-95.6%); specificity, 95.2% (95% CI, 86.7%-98.3%); PPV, 95.0%; NPV, 90.8%; LR+, 18.7; and LR,−0.10. Overall accuracy was 92.8%, with 9 of 125 knees misclassified (3 false positives, 6 false negatives).

Table 4

Model Performance and Calibration Metrics ^a

Metric	Value	Interpretation
Apparent AUC	0.962	Excellent
Optimism-corrected AUC	0.957	Excellent
Calibration-in-the-large	0	Good
Calibration slope	1	Very good
Brier score	0.069	Very good

AUC, area under the receiver operating characteristic curve. The composite model included tibial axis–condylar distance difference, lateral tibial height, and lateral femoral condylar index.

Figure 6.

Calibration plot showing agreement between predicted and observed anterior cruciate ligament rupture risk. C.L., confidence limits; LFCI, lateral femoral condylar index; LTH, lateral tibial height; T-ACD, tibial axis–condylar distance difference.

Figure 7.

Receiver operating characteristic (ROC) curve demonstrating excellent discriminative ability.

Table 5

Diagnostic Performance at Optimal Cutoff ^a

Metric	Value	95% CI
Optimal cutoff	0.471	–
Sensitivity, %	90.5	80.7%-95.6%
Specificity, %	95.2	86.7%-98.3%
Positive predictive value, %	95.0	–
Negative predictive value, %	90.8	–
Positive likelihood ratio	18.7	–
Negative likelihood ratio	0.10	–
Overall accuracy, %	92.8	–
Confusion matrix at optimal cutoff (0.471)
	ACL Intact (actual)	ACL Rupture (actual)	Total
Predicted intact	59 (TN)	6 (FN)	65
Predicted rupture	3 (FP)	57 (TP)	60
Total	62	63	125

The optimal cutoff was determined using Youden index. Dashes indicate not applicable (95% CIs not calculated for predictive values, likelihood ratios, and overall accuracy). ACL, anterior cruciate ligament; FN, false negative; FP, false positive; TN, true negative; TP, true positive.

Discussion

The major findings of our study demonstrated that coronal plane condylar height asymmetry is significantly associated with ACL rupture. Specifically, T-ACD was markedly greater in ACL-deficient knees than in controls (7.11 ± 1.96 mm vs 4.41 ± 1.65 mm; P < .001), with a T-ACD threshold of ≥4.5 mm conferring an odds ratio of 1.84/mm (95% CI, 1.34-2.53) and good discriminative ability (AUC = 0.858). Similarly, F-ACD (8.92 ± 2.18 mm vs 7.47 ± 2.09 mm; P < .001), LTH (4.87 ± 1.70 mm vs 2.76 ± 0.88 mm; P < .001), and LFCI (0.715 ± 0.056 vs 0.660 ± 0.049; P < .001) were all significantly greater in the rupture group, each with large effect sizes (Cohen d≥ 0.67). The composite model combining T-ACD, LTH, and LFCI achieved excellent discriminative accuracy (optimism-corrected AUC = 0.957; Nagelkerke R² = 0.83), with a sensitivity of 90.5% and specificity of 95.2% at the optimal cutoff. In contrast, PTS, NWI, and LFC-H/D showed no significant association with ACL rupture (P > .05).

The etiology of ACL rupture is multifactorial: sport-specific demands, playing surface properties and neuromuscular control all matter, yet anatomic variation is increasingly recognized as a key contributor.⁷ Among these anatomic factors, PTS, femoral condylar morphology, and tibial plateau geometry have received growing attention. It is important to note that this millimeter-level risk interpretation applies specifically to radiograph-based coronal measurements (T-ACD, F-ACD), which benefit from weightbearing standardization and reduced projection variability. MRI-based parameters (LTH, MTD, LFCI, LFC-H/D, NWI), while complementary, are subject to different sources of measurement uncertainty including slice selection variability and partial-volume effects, and their millimeter-scale thresholds should not be directly equated with radiographic measures without modality-specific validation. Conversely, an F-ACD cutoff around 8 mm offered only modest clinical sensitivity and specificity. Thus, our data suggest that the T-ACD is the more pertinent metric. These results reinforce the concept that coronal plane mismatches in tibial condylar height—rather than femoral differences alone—may play an important role in ACL biomechanics and injury vulnerability.

In our cohort, the F-ACD was significantly greater in ACL-deficient knees than in controls (8.9 ± 2.2 mm vs 7.5 ± 2.1 mm; P < .001). A larger F-ACD implies that the LFC sits more posteroinferior—or, put differently, farther from the femoral anatomic axis—thereby lengthening the functional span of the ACL when the knee is flexed.²³ This relative laxity can amplify anterior tibial translation and rotational instability during pivot-shift maneuvers, suggesting that a high F-ACD may serve as a coronal plane morphometric marker associated with ACL rupture risk. By the same rationale, the T-ACD captures the height disparity between the medial and lateral tibial plateaus. Prior work has linked an elevated or more convex lateral plateau to a pronounced pivot shift and secondary ACL insufficiency,¹² suggesting that an increased T-ACD represents a complementary morphologic marker potentially associated with augmented rupture risk.

Our demographic analysis showed that patients with ACL rupture were significantly younger than controls (27.3 ± 7.9 years vs 30.3 ± 7.6 years; P = .04), consistent with epidemiological reports that ACL injuries cluster in young, highly active populations.^15,18 Although the rupture group contained a larger proportion of male patients (74.6% vs 59.7%), the difference did not reach statistical significance (P = .08). This aligns with the mixed picture in the literature, where large registries report more absolute injuries in men, whereas sport-specific studies highlight a relatively higher noncontact risk in female athletes.^6,20 Mean BMI and BMI category distribution did not differ between groups; the absence of obese participants and the small number of overweight cases limited any assessment of BMI as an independent factor.^5,15 While some studies suggest that higher BMI increases ACL stress through greater mechanical load, more recent work indicates that obesity is not an isolated risk factor but may amplify risk when combined with PTS, notch stenosis, or other morphologic variables.⁵ Accordingly, the influence of BMI on ACL injury should be interpreted in conjunction with an individual's anatomic features and activity level. Finally, the similar side distribution (right vs left) supports that our morphometric measurements were not biased by limb dominance.

In our series, the LFCI was significantly higher in ACL-deficient knees than in controls (0.715 ± 0.056 vs 0.660 ± 0.049; P < .001). By expressing the AP length of the lateral condyle as a proportion of the total distal femoral AP length, the LFCI normalizes for interindividual bone size and captures the relative posterior positioning of the lateral condyle—a “global” femoral morphology that may more directly influence ACL loading by altering tibiofemoral contact kinematics during the pivot-shift maneuver.^9,13 ROC analysis yielded an AUC of 0.781, with an optimal cutoff of ≥0.698 (≈ 0.70) providing 69.8% sensitivity and 75% specificity. These findings echo earlier reports demonstrating a pronounced effect of femoral morphology on ACL biomechanics^9,17,19 and underscore the clinical importance of systematically assessing femoral, as well as tibial, geometric parameters when evaluating ACL injury risk.

The LFC-H/D did not differ significantly between groups in our cohort.⁷ Similarly, while MTD showed nominal significance, it demonstrated poor discriminative ability (AUC = 0.605) and lost significance in multivariate analysis. These null or weak findings may reflect the inherent measurement sensitivity of all MRI-derived parameters to slice selection and positioning, a limitation amplified by our 3-mm slice thickness. However, the robust associations observed for LTH (AUC = 0.854; large effect size) and LFCI (AUC = 0.781; large effect size) suggest that despite this technical constraint, parameters with sufficiently large effect sizes and high interobserver reliability can still demonstrate discriminative value. The discrepancy between parameters likely reflects both the magnitude of true anatomic differences between groups and the measurement precision achievable within our imaging protocol's constraints.

Interpretation of null findings requires careful consideration of whether they represent true absence of association versus methodological insensitivity. PTS showed no significant difference between groups, with both cohorts exhibiting mean values well below the established ≥12° risk threshold,²² suggesting a population-specific true negative finding. The NWI null finding aligns with recent large-scale MRI studies that question its discriminative value,^1,21 in contrast to earlier radiographic studies which suggested that narrower notches increase ACL injury risk.¹⁶ Our multivariate model, simultaneously adjusting for coronal plane tibial and femoral asymmetry, likely captured the biomechanical variance previously attributed to notch stenosis in univariate analyses. In contrast, LFC-H/D and MTD findings remain ambiguous and may represent methodological false negatives. Grassi et al⁷ identified deep posterior LFCs as a significant risk factor in multiple ACL failures, while our 3-mm slice thickness may have insufficient resolution to capture this subtle geometry. Similarly, literature on MTD remains inconsistent: Hashemi et al⁸ reported significant associations, whereas Akhavi Milani et al² found none using conventional MRI. Both parameters depend critically on precise subchondral contour identification, measurements highly vulnerable to partial-volume averaging. Future investigations employing high-resolution volumetric MRI (isotropic ≤1.5 mm) or computed tomography–based morphometry are needed to clarify whether these represent true null findings or methodological artifacts.

Lateral tibial compartment geometry showed stronger associations with ACL rupture than medial compartment features. LTH was markedly greater in ACL-deficient knees (4.87 ± 1.70 mm vs 2.76 ± 0.88 mm; P < .001) and demonstrated strong discriminative power (AUC = 0.854), remaining an independent predictor with an odds ratio ≈4 even after multivariate adjustment. A higher lateral plateau rim may reduce femoral contact area during dynamic loading and amplify rotational instability.⁷ This lateral-predominant pattern aligns with biomechanical models emphasizing lateral compartment geometry in ACL loading.¹⁰

PTS is one of the most frequently investigated morphological risk factors for ACL rupture, as a steeper slope augments anterior tibial translation and increases ligament load.¹¹ Clinical series report a sharp rise in injury risk once PTS approaches ≥12°, with odds escalating 5- to 11-fold.²² In our cohort, however, mean PTS values remained well below this threshold in both groups (10.9°± 3.2° vs 10.7°± 3.0°; P = .68), consistent with a population-specific true negative finding.

Our data indicate that T-ACD ≥4.5 mm or F-ACD ≥8 mm mark thresholds beyond which coronal alignment becomes structurally imbalanced, substantially increasing ACL injury likelihood. At-risk patients—particularly those engaged in cutting or pivoting sports—should be counseled to adopt targeted neuromuscular training programs, including quadriceps-to-hamstring strengthening and proprioceptive drills, to mitigate the excessive mechanical load potentially associated with tibiofemoral mismatch. These morphometric thresholds may also be relevant in postoperative settings, where patients with high-risk anatomy undergoing ACL reconstruction may benefit from enhanced rehabilitation protocols to reduce retear risk.

Our comprehensive internal validation demonstrated that the composite prediction model provides both excellent discrimination (optimism-corrected AUC = 0.957) and well-calibrated risk estimates, with minimal optimism (0.005) and ideal calibration metrics (calibration-in-the-large = 0.000; calibration slope = 1.000; Brier score = 0.069). At the optimal cutoff, the positive likelihood ratio of 18.7 indicates that athletes with high-risk morphometry warrant targeted prevention programs, while the negative likelihood ratio of 0.10 provides reassurance for those with normal profiles. The high specificity (95.2%) further minimizes unnecessary interventions in clinical practice.

Limitations

This investigation employed a retrospective case-control design, which precludes causal inference; it can only characterize associations between ACL rupture and morphometric parameters. Because the study population consisted of patients who underwent MRI after acute knee trauma, selection bias is possible and the findings may not be generalizable to the broader population. Most morphometric measurements were obtained from 2-dimensional radiographic or MRI slices; without 3-dimensional reconstruction, the full spatial complexity of bony anatomy may not have been captured.

A notable methodological constraint is the 3-mm MRI slice thickness employed in this study. Several morphometric parameters—particularly LFC-H/D, LTH, MTD, and NWI—were derived from single-slice measurements, introducing inherent partial-volume effects and through-plane spatial uncertainty. Although our protocol mandated selection of the slice offering the clearest anatomic landmarks and both observers underwent joint training to standardize slice identification, the 3-mm acquisition resolution means that small shifts in slice position could theoretically alter measurements by magnitudes approaching the reported group differences. This technical limitation may attenuate effect sizes and reduce discriminative power, particularly for parameters showing modest group differences (eg, MTD). The high inter- and intraobserver reliability (ICC > 0.80 for all parameters) suggests that, within the constraints of our imaging protocol, measurements were consistently reproducible; however, future investigations should employ high-resolution imaging protocols—specifically thin-slice MRI with isotropic voxel acquisition (≤1.5 mm) or 3-dimensional volumetric sequences with multiplanar reconstruction—to validate the proposed thresholds (T-ACD ≥4.5 mm, F-ACD ≥8 mm) and reduce projection- and slice-related measurement uncertainty. Such advanced imaging would provide superior spatial resolution, eliminate partial-volume averaging, and enable true 3-dimensional morphometric analysis, potentially revealing more precise associations and clarifying ambiguous findings for parameters such as LFC-H/D and MTD. Until such higher-resolution validation is performed, our millimeter-scale findings should be interpreted with appropriate caution regarding their absolute precision.

Potential confounders—such as sport-specific exposure, meniscal integrity, generalized ligamentous laxity and neuromuscular control—were not systematically recorded and therefore could not be included in the analyses. Although radiographic acquisition followed a standard protocol (weightbearing, near-full extension, patellae forward, standardized beam alignment), T-ACD and F-ACD are inherently projection-dependent measurements that may be influenced by small variations in limb positioning at the time of image acquisition. Specifically, tibial or femoral rotation, flexion-extension mismatch, subtle varus-valgus alignment changes, or beam-centering variations could meaningfully alter measured condylar distances. While our high interobserver reliability (ICC > 0.90 for both parameters) suggests that measurements were reproducible within a given projection, we did not assess repeatability across multiple radiographic projections or quantify the sensitivity of T-ACD and F-ACD values to controlled alterations in positioning. This projection sensitivity is particularly relevant given our emphasis on millimeter-scale thresholds (T-ACD ≥ 4.5 mm; F-ACD ≥ 8 mm) for risk stratification. Future studies should evaluate the robustness of these measurements across standardized variations in radiographic technique (eg, ±5° rotation, ±3° flexion) to establish the minimal clinically important difference and inform threshold stability. Until such validation is performed, clinicians should interpret these cutoffs with recognition that projection-related measurement variability may approach 1 to 2 mm even with careful technique. Finally, the study was conducted at a single center with limited ethnic and morphological diversity; external validation in larger, multicenter cohorts is warranted to confirm the applicability of these results to other populations.

Conclusion

Footnotes

Final revision submitted March 11, 2026; accepted March 20, 2026.

The authors declared that there are no conflicts of interest in the authorship and publication of this contribution. AOSSM checks author disclosures against the Open Payments Database (OPD). AOSSM has not conducted an independent investigation on the OPD and disclaims any liability or responsibility relating thereto.

Ethical approval for this study was obtained from the ethics committee of Şişli Hamidiye Etfal Training and Research Hospital (No. 4809).

ORCID iDs

Enver İpek

Mehmet Akif Köse

Abdurrahman Demirhan

Osman Tuğrul Eren

References

Akcaalan

Kavaklilar

Caglar

Ugurlu

Dogan

Akkaya

Investigation of morphometric factors associated with adolescent ACL rupture. Orthop J Sports Med. 2023;11(9):23259671231194928.

Akhavi Milani

Motallebzadeh Khanamiri

Mahmoudpour

. The geometric features of tibiofemoral joint associated with anterior cruciate ligament injury: reliability of measurements on conventional MRI. Eur J Radiol Open. 2022;9:100420. doi:10.1016/j.ejro.2022.100420

Arundale

AJH

Silvers-Granelli

Myklebust

. ACL injury prevention: where have we come from and where are we going? J Orthop Res. 2022;40(1):43-54. doi:10.1002/JOR.25058

Bayer

Meredith

Wilson

, et al. Knee morphological risk factors for anterior cruciate ligament injury: a systematic review. J Bone Joint Surg Am. 2020;102(8):703-718. doi:10.2106/JBJS.19.00535

Bojicic

Beaulieu

Imaizumi Krieger

Ashton-Miller

Wojtys

EM.

Association between lateral posterior tibial slope, body mass index, and ACL injury risk. Orthop J Sports Med. 2017;5(2):2325967116688664.

Dewig

Boltz

Moffit

Rao

Collins

Chandran

Epidemiology of anterior cruciate ligament tears in National Collegiate Athletic Association athletes: 2014/2015-2018/2019. Med Sci Sports Exerc. 2024;56(1):29-36.

Grassi

Macchiarola

Urrizola Barrientos

, et al. Steep posterior tibial slope, anterior tibial subluxation, deep posterior lateral femoral condyle, and meniscal deficiency are common findings in multiple anterior cruciate ligament failures: an MRI case-control study. Am J Sports Med. 2019;47(2):285-295. doi:10.1177/0363546518823544

Hashemi

Chandrashekar

Mansouri

, et al. Shallow medial tibial plateau and steep medial and lateral tibial slopes: new risk factors for anterior cruciate ligament injuries. Am J Sports Med. 2010;38(1):54-62. doi:10.1177/0363546509349055

Hodel

Kabelitz

Tondelli

Vlachopoulos

Sutter

Fucentese

SF.

Introducing the lateral femoral condyle index as a risk factor for anterior cruciate ligament injury. Am J Sports Med. 2019;47(10):2420-2426. doi:10.1177/0363546519858612

10.

Kaarre

Ahrendt

Giusto

Özbek

Apseloff

Musahl

The effect of bony morphology on anterior cruciate ligament injury and surgery. Acta Orthop Traumatol Turc. 2024;58(2):83-88.

11.

Koyejo

Babalola

Itakpe

Maleeq

RO.

Evaluation of increased posterior tibial slope as a risk factor for anterior cruciate ligament injury. J Glob Med. 2024;4(suppl 1):e165. doi:10.51496/JOGM.V4.S1.165

12.

Kujala

Nelimarkka

Koskinen

SK.

Relationship between the pivot shift and the configuration of the lateral tibial plateau. Arch Orthop Trauma Surg. 1992;111(4):228-229. doi:10.1007/BF00571483

13.

Zhang

Chen

Fang

The relationship between lateral femoral condyle index and noncontact anterior cruciate ligament rupture. Int J Gen Med. 2022;15:1789. doi:10.2147/IJGM.S347742

14.

Mandrekar

JN.

Receiver operating characteristic curve in diagnostic test assessment. J Thorac Oncol. 2010;5(9):1315-1316. doi:10.1097/JTO.0b013e3181ec173d

15.

Motififard

Akbari Aghdam

Ravanbod

, et al. Demographic and injury characteristics as potential risk factors for anterior cruciate ligament injuries: a multicentric cross-sectional study. J Clin Med. 2024;13(17):5063. doi:10.3390/JCM13175063

16.

Pękala

Podgórski

Shukla

Winnicka

Biernacka

Grzelak

. Do variants of the intercondylar notch predispose children to the injury of the anterior cruciate ligament? Clin Anat. 2019;32(5):706-709. doi:10.1002/CA.23382

17.

Pfeiffer

Burnham

Hughes

, et al. An increased lateral femoral condyle ratio is a risk factor for anterior cruciate ligament injury. J Bone Joint Surg Am. 2018;100(10):857-864. doi:10.2106/JBJS.17.01011

18.

Sanders

Maradit Kremers

Bryan

, et al. Incidence of anterior cruciate ligament tears and reconstruction: a 21-year population-based study. Am J Sports Med. 2016;44(6):1502-1507. doi:10.1177/0363546516629944

19.

Sun

Tang

The relationship between lateral femoral condyle ratio measured by MRI and anterior cruciate ligament injury. Front Bioeng Biotechnol. 2024;12:1362110. doi:10.3389/FBIOE.2024.1362110

20.

Sutton

Bullock

JM.

Anterior cruciate ligament rupture: differences between males and females. J Am Acad Orthop Surg. 2013;21(1):41-50.

21.

van Eck

Kopf

van Dijk

Tashman

Comparison of 3-dimensional notch volume between subjects with and subjects without anterior cruciate ligament rupture. Arthroscopy. 2011;27(9):1235-1241.

22.

Webb

Salmon

Leclerc

Pinczewski

Roe

JP.

Posterior tibial slope and further anterior cruciate ligament injuries in the anterior cruciate ligament–reconstructed patient. Am J Sports Med. 2013;41(12):2800-2804. doi:10.1177/0363546513503288

23.

Xie

Teng

, et al. Increased lateral femoral condyle ratio is associated with a greater risk of anterior cruciate ligament injury and concomitant anterolateral ligament and meniscus injuries. Arthroscopy. 2025;41(8):3255-3267.