A Diasus® diagnostic musculoskeletal ultrasound (US) scanner (System 8, Dynamic imaging, Livingston, Scotland, UK), was used for both studies one and two (figure 21a). Scanning was completed in B-Mode to provide real-time grey-scale images, sampled at a maximum frame rate of 30 frames per second. The return echo signals were automatically processed using Diasus® 2D spline filtering. Image pixilation was standardised at 640 x 440 pixels, the optimum settings for fine image resolution available using this software. The overall transmit power and gain was set at ≤50 and ≤30 respectively, in accordance with the European League Against Rheumatism (EULAR) working group for US in rheumatology scanning recommendations (Backhaus 2001). However, grey-scale contrast was continually adjusted during image acquisition in real-time using multiple fine gain control and focus points. Where possible the least amount of focus points were used and centred at the intermetatarsal level for plantar foot scans and the upper third of the joint space for dorsal foot scans. This enabled the transmit frequency to be as high as possible to achieve good image resolution whilst also maintaining a suitable wave
penetration depth aimed at the level of anatomical interest.
All US scanning was performed in accordance with the British Medical Ultrasound Society (BMUS) guidelines for safe use (Fitzpatrick et al. 1998). In addition, the image acquisition protocol was designed to reflect the ALARA principles (as low as reasonably achievable) reported by the American Institute of Ultrasound in Medicine (AIUM), (Fitzpatrick et al. 1998). Thus, the minimum amount of US exposure was used to reasonably fulfil the objectives of the US scan.
Figure 21: US equipment
Where 21a illustrates the Diasus® portable US unit, system 8, 21b illustrates the 5-10MHz ultra wideband linear-array transducer, active length 40mm (left) and the 8-16MHz transducer, active length 26mm (right).
Images author’s own.
The dual probe system operates with two linear array transducers (figure 21b). This enabled specific sound wave frequency use and thus optimised image resolution where possible whilst also ensuring accurate wave penetration depth when required, as demonstrated in figure 22. For example, the 8-16MHz transducer was not sufficient to accurately review the intermetatarsal spaces at the level of the deep transverse intermetatarsal ligament and therefore the 5-10MHz transducer was used for these scans.
Figure 22: US transducer frequency & tissue depth penetration
A Longitudinal skin section demonstrating US frequency compared to depth of tissue penetrated in a large joint. Image reproduced with permission from RH09 schematic design, Southampton (2009)
3.6.3.1 US protocol
The US foot scan was completed prior to the podiatric assessment or evaluation of disease state, to minimise the potential for observer bias; the researcher completed the scan without prior knowledge of the participant’s foot health or disease activity status. An overview of the US
scanning protocol is shown in figure 23. Hypo-allergenic, alcohol free coupling gel was liberally used throughout to improve transducer to skin contact.
Figure 23: US scanning protocol
1. Participant is seated on a flatbed plinth, with their feet facing towards the researcher
2. The transducer is applied to the plantar forefoot region, at the level of the first metatarsal head, orientated in the transverse plane
3. Structural landmarks are identified (sesamoid bones) for anatomical orientation
4.The US scan is completed, using the 5-10MHz ultra wideband linear array transducer, moving proximally to distally in this region
5. The transducer is sequentially relocated medially to laterally, with proximal to distal scan sequences repeated. The central portion of the transducer is positioned over the MTP joint region.
6. The transducer is applied to the plantar forefoot region, at the level of the first metatarsal head, orientated in the longitudinal plane
7. Structural landmarks are identified (base of metatarsal head and proximal phalanx) for anatomical orientation. The centre of the transducer is aligned with the MTP joint space.
8. The US scan is completed, moving medially to laterally in this region
9. The longitudinal plantar scanning sequence is repeated plantar to all MTP joint regions and intermetatarsal regions at the level of the MTP joint
9b. Observed plantar lesions must be scanned in both the transverse and longitudinal planes before a positive identification is recorded
10. The transducer is applied to the dorsal forefoot region, at the level of the first MTP joint, orientated in the longitudinal plane
Forefoot bursae were noted as present if detectable in both the transverse and longitudinal planes, when scanning from a plantar approach, as illustrated in figures 24a and 24b. There are no standardised documented approaches for the determination of intermetatarsal or plantar bursae in the forefoot. However, common differential diagnoses include intermetatarsal neuroma and flexor digitorum longus tenosynovitis, for which a plantar US approach is recommended (Backhaus 2001, Baker et al., Brown 2005, Fitzpatrick et al. 1998, Koski 1998, Chauveaux et al. 1987). The proposed plantar approach is consistent with that used by Bowen et al. (Bowen et al. 2008, Bowen et al.), who demonstrated reliable detection of FFB in the baseline and year- one follow-up studies.
Figure 24: US transducer orientations
Where 24a illustrates the plantar transverse scan at the level of the MTP joint region, 24b illustrates the plantar longitudinal scan at the level of the second MTP joint region, and 24c illustrates the dorsal longitudinal scan at the level of the third MTP joint region. Images author’s own.
MTP joint synovitis was noted as present if detectable in both the transverse and longitudinal planes when scanning from a dorsal approach (figure 24c). The selected approach conforms to those proposed by the EULAR working group for US in rheumatology (Backhaus 2001). Metatarsal head erosion was noted as present if detectable in either the dorsal or plantar scanning approach. However, a positive annotation was only given if the erosion was detectable in both the transverse and longitudinal plane, in accordance with EULAR guidelines (Backhaus 2001). Findings were recorded on the US assessment form (see appendix section A7c). 3.6.3.2 Benefits & limitations of US
Real-time multi-planar grey-scale US, in B-Mode, allows accurate detection of bone and soft tissue lesions within the forefoot. The use of Power Doppler would provide additional benefit for the identification of active inflammation. However, this Power Doppler was not available in this study. Image artefact, particularly anisotropy (disparity in acoustic feedback with changes in transducer orientation) was problematic when scanning plantarly due to the large number of converging, differently orientated, anatomical structures. To overcome this, the transducer was applied perpendicularly to the sole of the foot and then angled over a range of -45° to +45° about this original 90° position thus altering acoustic enhancement. The use of positional acoustic variation over striated tendonous structures provided further clarification regarding tissue detection and differentiation. Where fluid filled cavities were detected, the transducer was
held in a still position for a minimum of 5 seconds to observe any potential blood vessel pulsation. Gentle pressure was applied to the transducer to compress observed fluid to identify capsulation or distribution.
For dorsal MTP joint scanning, good transducer to skin contact was often difficult due to the presence of forefoot deformity, particularly lesser digit retraction or subluxation. The use of a smaller ‘hockey-stick’ transducer may have improved image acquisition in this area by improving transducer to skin coupling (Backhaus 2001). However, this transducer was not available in this study. Stand-off pad use was trialled prior to data collection in order to improve transducer to skin coupling over the lesser digits (Warner et al. 2008, Brown 2005, Riente et al. 2006). However the frequency reduction required to image at sufficient depth when using a stand-off pad noticeably reduced image quality. Thus, the smaller linear 8-16MHz transducer was preferentially chosen.