Normal Values of Distal Radioulnar Joint Kinematics during a Dynamic Press Test

Background  Measurement of in vivo distal radioulnar joint (DRUJ) pathomechanics during simple activities can represent the disability experienced by patients and may be useful in diagnostics of DRUJ instability. A first step is to describe the physiological normal limits for DRUJ kinematics in a reproducible and precise test setup, which was the aim of this study. Methods  DRUJ kinematics were evaluated in 33 participants with dynamic radiostereometry (RSA) while performing a standardized press test examination. AutoRSA software was used for image analyses. Computed tomography (CT) forearm bone models were generated, and standardized anatomical axes were applied to estimate kinematic outcomes including, DRUJ translation, DRUJ position ratio, and changes in ulnar variance. Repeatability of dynamic RSA press test double examinations was evaluated to estimate the precision and intraclass correlation coefficient (ICC) test-retest agreement. Results  The maximum force during the press test was 6.0 kg (95% confidence interval [CI]: 5.1-6.9), which resulted in 4.7 mm (95% CI: 4.2-5.1) DRUJ translation, DRUJ position ratio of 0.40 (95% CI: 0.33-0.44), and increase in ulnar variance of 1.1 mm (95% CI: 1.0-1.2). The mean maximum DRUJ translation leveled off after a 5 kg force application. The DRUJ translation ICC coefficient was 0.93 within a prediction interval of ± 0.53mm. Conclusions  This clinical study demonstrates the normal values of DRUJ kinematics and reports excellent agreement and high precision of the press tests examination using an automated noninvasive dynamic RSA imaging method based on patient-specific CT bone models. The next step is the application of the method in patients with arthroscopic verified triangular fibrocartilage complex injuries. Level of Evidence  This is a Level IV, case series study.

A clinical decision support system optimising adjuvant chemotherapy for colorectal cancers by integrating deep learning and pathological staging markers: a development and validation study

BACKGROUND

The DoMore-v1-CRC marker was recently developed using deep learning and conventional haematoxylin and eosin-stained tissue sections, and was observed to outperform established molecular and morphological markers of patient outcome after primary colorectal cancer resection. The aim of the present study was to develop a clinical decision support system based on DoMore-v1-CRC and pathological staging markers to facilitate individualised selection of adjuvant treatment.

METHODS

We estimated cancer-specific survival in subgroups formed by pathological tumour stage (pT<4 or pT4), pathological nodal stage (pN0, pN1, or pN2), number of lymph nodes sampled (≤12 or >12) if not pN2, and DoMore-v1-CRC classification (good, uncertain, or poor prognosis) in 997 patients with stage II or III colorectal cancer considered to have no residual tumour (R0) from two community-based cohorts in Norway and the UK, and used these data to define three risk groups. An external cohort of 1075 patients with stage II or III R0 colorectal cancer from the QUASAR 2 trial was used for validation; these patients were treated with single-agent capecitabine. The proposed risk stratification system was evaluated using Cox regression analysis. We similarly evaluated a risk stratification system intended to reflect current guidelines and clinical practice. The primary outcome was cancer-specific survival.

FINDINGS

The new risk stratification system provided a hazard ratio of 10·71 (95% CI 6·39-17·93; p<0·0001) for high-risk versus low-risk patients and 3·06 (1·73-5·42; p=0·0001) for intermediate versus low risk in the primary analysis of the validation cohort. Estimated 3-year cancer-specific survival was 97·2% (95% CI 95·1-98·4; n=445 [41%]) for the low-risk group, 94·8% (91·7-96·7; n=339 [32%]) for the intermediate-risk group, and 77·6% (72·1-82·1; n=291 [27%]) for the high-risk group. The guideline-based risk grouping was observed to be less prognostic and informative (the low-risk group comprised only 142 [13%] of the 1075 patients).

INTERPRETATION

Integrating DoMore-v1-CRC and pathological staging markers provided a clinical decision support system that risk stratifies more accurately than its constituent elements, and identifies substantially more patients with stage II and III colorectal cancer with similarly good prognosis as the low-risk group in current guidelines. Avoiding adjuvant chemotherapy in these patients might be safe, and could reduce morbidity, mortality, and treatment costs.

FUNDING

The Research Council of Norway.

Designing deep learning studies in cancer diagnostics

The number of publications on deep learning for cancer diagnostics is rapidly increasing, and systems are frequently claimed to perform comparable with or better than clinicians. However, few systems have yet demonstrated real-world medical utility. In this Perspective, we discuss reasons for the moderate progress and describe remedies designed to facilitate transition to the clinic. Recent, presumably influential, deep learning studies in cancer diagnostics, of which the vast majority used images as input to the system, are evaluated to reveal the status of the field. By manipulating real data, we then exemplify that much and varied training data facilitate the generalizability of neural networks and thus the ability to use them clinically. To reduce the risk of biased performance estimation of deep learning systems, we advocate evaluation in external cohorts and strongly advise that the planned analyses, including a predefined primary analysis, are described in a protocol preferentially stored in an online repository. Recommended protocol items should be established for the field, and we present our suggestions.

Assessment of knee kinematics with dynamic radiostereometry: Validation of an automated model-based method of analysis using bone models

Radiostereometic analysis (RSA) is a precise method for the functional assessment of joint kinematics. Traditionally, the method is based on tracking of surgically implanted bone markers and analysis is user intensive. We propose an automated method of analysis based on models generated from computed tomography (CT) scans and digitally reconstructed radiographs. The study investigates method agreement between marker-based RSA and the CT bone model-based RSA method for assessment of knee joint kinematics in an experimental setup. Eight cadaveric specimens were prepared with bone markers and bone volume models were generated from CT-scans. Using a mobile fixture setup, dynamic RSA recordings were obtained during a knee flexion exercise in two unique radiographic setups, uniplanar and biplanar. The method agreement between marker-based and CT bone model-based RSA methods was compared using bias and LoA. Results obtained from uniplanar and biplanar recordings were compared and the influence of radiographic setup was considered for clinical relevance. The automated method had a bias of -0.19 mm and 0.11° and LoA within ±0.42 mm and ±0.33° for knee joint translations and rotations, respectively. The model pose estimation of the tibial bone was more precise than the femoral bone. The radiographic setup had no clinically relevant effect on results. In conclusion, the automated CT bone model-based RSA method had a clinical precision comparable to that of marker-based RSA. The automated method is non-invasive, fast, and clinically applicable for functional assessment of knee kinematics and pathomechanics in patients.

Distal radioulnar joint stability measured with radiostereometry during the piano key test

Distal radioulnar joint instability is difficult to grade by clinical examination and interobserver reliability is low. This study used a new and precise radiostereometry method for measurement of distal radioulnar joint translation. Eight human donor arms were positioned in a custom-made fixture and a standardized piano key test was done with pressure on the ulnar head. Examination was done before and after dividing the styloid and foveal insertions of the triangular fibrocartilage complex. In the intact wrists, the piano key test induced a mean 1.36 mm translation of the ulnar head, which increased statistically significantly to 1.96 mm after a lesion of the styloid ligament insertion and to 2.3 mm after combined lesions of the styloid and foveal ligament insertions. This experimental cadaver study demonstrates a radiological method for precise quantification of distal radioulnar joint stability after different grades of triangular fibrocartilage complex injury.

Deep learning for prediction of colorectal cancer outcome: a discovery and validation study

BACKGROUND

Improved markers of prognosis are needed to stratify patients with early-stage colorectal cancer to refine selection of adjuvant therapy. The aim of the present study was to develop a biomarker of patient outcome after primary colorectal cancer resection by directly analysing scanned conventional haematoxylin and eosin stained sections using deep learning.

METHODS

More than 12 000 000 image tiles from patients with a distinctly good or poor disease outcome from four cohorts were used to train a total of ten convolutional neural networks, purpose-built for classifying supersized heterogeneous images. A prognostic biomarker integrating the ten networks was determined using patients with a non-distinct outcome. The marker was tested on 920 patients with slides prepared in the UK, and then independently validated according to a predefined protocol in 1122 patients treated with single-agent capecitabine using slides prepared in Norway. All cohorts included only patients with resectable tumours, and a formalin-fixed, paraffin-embedded tumour tissue block available for analysis. The primary outcome was cancer-specific survival.

FINDINGS

828 patients from four cohorts had a distinct outcome and were used as a training cohort to obtain clear ground truth. 1645 patients had a non-distinct outcome and were used for tuning. The biomarker provided a hazard ratio for poor versus good prognosis of 3·84 (95% CI 2·72-5·43; p<0·0001) in the primary analysis of the validation cohort, and 3·04 (2·07-4·47; p<0·0001) after adjusting for established prognostic markers significant in univariable analyses of the same cohort, which were pN stage, pT stage, lymphatic invasion, and venous vascular invasion.

INTERPRETATION

A clinically useful prognostic marker was developed using deep learning allied to digital scanning of conventional haematoxylin and eosin stained tumour tissue sections. The assay has been extensively evaluated in large, independent patient populations, correlates with and outperforms established molecular and morphological prognostic markers, and gives consistent results across tumour and nodal stage. The biomarker stratified stage II and III patients into sufficiently distinct prognostic groups that potentially could be used to guide selection of adjuvant treatment by avoiding therapy in very low risk groups and identifying patients who would benefit from more intensive treatment regimes.

FUNDING

The Research Council of Norway.

Reliability of computer-assisted periacetabular osteotomy using a minimally invasive approach

BACKGROUND

Periacetabular osteotomy (PAO) is the treatment of choice for younger patients with developmental hip dysplasia. The procedure aims to normalize the joint configuration, reduce the peak-pressure, and delay the development of osteoarthritis. The procedure is technically demanding and no previous study has validated the use of computer navigation with a minimally invasive transsartorial approach.

METHODS

Computer-assisted PAO was performed on ten patients. Patients underwent pre- and postoperative computed tomography (CT) scanning with a standardized protocol. Preoperative preparation consisted of outlining the lunate surface and segmenting the pelvis and femur from CT data. The Biomechanical Guidance System was used intra-operatively to automatically calculate diagnostic angles and peak-pressure measurements. Manual diagnostic angle measurements were performed based on pre- and postoperative CT. Differences in angle measurements were investigated with summary statistics, intraclass correlation coefficient, and Bland-Altman plots. The percentage postoperative change in peak-pressure was calculated.

RESULTS

Intra-operative reported angle measurements show a good agreement with manual angle measurements with intraclass correlation coefficient between 0.94 and 0.98. Computer navigation reported angle measurements were significantly higher for the posterior sector angle ([Formula: see text], [Formula: see text]) and the acetabular anteversion angle ([Formula: see text], [Formula: see text]). No significant difference was found for the center-edge ([Formula: see text]), acetabular index ([Formula: see text]), and anterior sector angle ([Formula: see text]). Peak-pressure after PAO decreased by a mean of 13% and was significantly different ([Formula: see text]).

CONCLUSIONS

We found that computer navigation can reliably be used with a minimally invasive transsartorial approach PAO. Angle measurements generally agree with manual measurements and peak-pressure was shown to decrease postoperatively. With further development, the system will become a valuable tool in the operating room for both experienced and less experienced surgeons performing PAO. Further studies with a larger cohort and follow-up will allow us to investigate the association with peak-pressure and postoperative outcome and pave the way to clinical introduction.