Translational research in diagnosis and management of soft tissue tumours

US plays an important role in detecting soft tissues masses, and with the use of Power Doppler and the recent introduction of echographic contrast media it well reveals the presence of intrinsic flow within the mass and its characteristics improving the Doppler signal. US specificity is not high, in some cases it is specific in confirming the hypothesis of a benign lesion like Baker cyst, ganglion, angioma, lypoma and ossifying myositis. In children sonography is in most cases the first imaging evaluation, and this is particularly useful for small and superficial lesions. It is a real time examination that can be carried out without problems of motion artifacts, and this is of utmost importance in children. Furthermore, dynamic scans can be helpful in the differential diagnosis (for example between STS and muscular hernias) [39]. Elastosonography is a new frontier of US, it allows to differentiate the grade of elasticity of a soft tissue mass compared to the adjacent tissues by a color map related to the compressibility of the mass. When elasticity is soft the lesion is generally benign, when less-soft it is borderline and when no elasticity is seen the lesion is malignant [40]. US elastosonography, together with B-mode, power and color Doppler can be a useful guide not only for biopsy but also in the follow-up of patients with soft tumours [41] (Fig. 3a-d).

Having almost replaced CT, MRI is the method of choice for detection, characterization and follow-up of soft tissues masses. Contrast in soft tissues is of higher quality allowing an easier detection of the lesion, and improves delineation of their extent and involvement of neurovascular structures and medullary bone [41]. Many STS have a non specific behaviour in T1 and T2-weighted images. A correct diagnosis can be made only by evaluating signal intensity, intrinsic lesion features such as site, size, and growth pattern. Contrast media may help differentiate cystic and necrotic components from those of solid tumour as well as assessing lesion aggressiveness. Today in literature there is no concordance about how traditional MR imaging can precisely differentiate benign from malignant lesions [42, 43] (Fig. 4a - d).

As for bone tumours, newer MRI techniques such as DCE, DWI, MR Spectroscopy MRS can be used not only to cope with the problem of differentiating between benign and malignant tumours but also to improve the possibility of a correct diagnosis [44, 45]. Quantitative DCE or DCE is a non invasive technique that estimates the percentage of necrosis in malignant tumours by comparing pre- and post- treatment examination. This new technique can identify not only post treatment necrosis but also early changes during treatment to modify ineffective therapies. In STS voxels of the necrotic area enhance slowly compared to those of the remaining viable area. In DCE, although the estimate of necrosis is obtained from the whole tumour volume, results are reliable and superimposable to histopathological evaluation obtained from micron thick sections of the resected specimen [46]. DCE is also used to differentiate benign from malignant soft tissue lesions [47]. Malignant lesions usually reveal an increased rate of enhancement of vascularity compared to the benign with a DCE ratio of 3:1 or 4:1 at the first passage [45, 48]. However, DCE poses a problem with highly cellular benign tumours that overlap with STS [44].

DWI is an MR technique that is sensitive to the random motion of water. Enhancement identifies areas of inflammatory post-surgical change as well as recurrence of disease. Unlike morphological MR imaging sequences and DCE, adds functional information about tissue composition without intravenous contrast [46]. DWI techniques depict tissues with signal intensity that vary in proportion to the average distance by which water molecules are displaced per unit time through the processes of water self-diffusion. DWI is based on the principle that diffusion of the water molecules produces a net dephasing of the spinning protons within a voxel resulting in a reduced signal intensity and image brightness. When examining volume by DWI, signal intensity decreases as the average speed of water, which travels along a temporarily generated magnetic field gradient increases. The quantity of water molecules displaced is described by the diffusion constant which varies with the direction. When the distance of water displacement is the same in all directions diffusion is isotropic. DW images are direction independent, that is, they are independent from the direction along which the magnetic field is applied. In the human body this occurs in fluids where mobile water molecules are free to move in all directions like CSF, brain ventricules or fluids in cystic lesions. The diffusion constant and average water displacement in these fluids are high and equal in all directions. The result: strong signal attenuation (dark) on DWI. In structures where the movement of water molecules is restricted by body structures diffusion is not the same in all directions, but is direction dependent or anisotropic. For example, brain water molecules diffuse faster along the axon with intact myelin sheaths than perpendicular to them. In other words diffusion signal loss is higher when the gradient is applied along the axons axis and lower when applied perpendicular. This technique was first used to detect acute brain ischemia, but during the last decade it has become an imaging biomarker for oncology applied to the entire body. Today DWI is used to identify and detect tissues where pathologic processes have altered the motion of fluids outside vessels and capillaries. However, unfortunately, there is no evidence relating imaging to genetic mutations, VEGF expression, KI 67 expression, and to several other relevant markers.

DWI is acquired with an echo-planar imaging sequence. Dephasing is produced by applying addition diffusion sensitizing gradients during the image acquisition cycle. Two magnetic field gradients with the same polarity are delivered between the excitation pulse and signal collection to sensibilize the sequence to diffusion effect. In the middle of the two gradients a 180° RF pulse is delivered to change the spin phase along the field gradient direction, thus all spins involved and displaced in the diffusion process, experience a different phase shift. Protons in water molecules that have moved will be found in a different position and field strength during the second gradient. These will not be completely rephrased and their magnetic moment will no longer add thus leading to a loss of signal. It is this reduction in signal intensity that produces a difference in contrast between the moving molecules (loss of signal intensity: dark) and not moving molecules (high signal intensity: bright). Signal attenuation depends on strength and duration of the gradient pulses, their temporal separation and the diffusion constant along the direction of the gradient field. The so called b-value quantifies the amount of signal loss with a given pulse sequence and for a given diffusion constant i.e. how sensitive a sequence is to diffusion effects. The diffusion constant in biological tissues can be measured by repeated scanning with different b-values but identical parameters, in particular unchanged gradient direction. Observed diffusion constants are indicated like Apparent Diffusion Coefficents (ADCs) to differentiate them from the constant of unrestricted diffusion in pure water. Using ADCs the so called ADC maps can be built: a grey scale represents the mean ADC of the corresponding voxel. It is important to note that an area of viable tumour bright on a DWI image (for reduced water mobility both for high cellularity and membrane integrity), will be dark on the corresponding ADC map (for its lower diffusion constant). Diffusion of water is in fact more restricted in tumours than in normal tissues and this on DWI is seen as a high signal intensity in viable tumours. DWI and ADC maps provide qualitative and quantitative information about tissue cellularity and cell integrity. Potentially this is helpful in identifying not only benign from malignant lesions but also in revealing necrotic tumours and peri tumoural edemas from residual viable tumours underscoring the efficacy of tumour response to therapy. DWI is highly sensitive to motion, in brain imaging particularly to rotation or trembling of the head, in trunk imaging to respiratory motion. To cope with these drawbacks DWI uses a single shot or multi shot Echo-planar imaging (EPI). It is fast and renders this techniques less sensitive to patient motion with the advantages of covering a large volume, a high signal to noise ratio and a low power deposition in tissues because several echoes are acquired after a single excitation pulse. Multishot EPI reduces the susceptibility artifacts although it increases sensitivity to motion and scan time, while single shot fast spin-echo (RARE or HASTE) are less sensitive to susceptibility but have more limitations in signal to noise ratio and blurring. Sometimes in clinical practice differentiation of benign or malignant tumours only on ADC quantification is not so easy, while non myxoid malignant tumours show significantly lower ADC values than the benign. In myxoid tumours the differentiation is not so clear for the long T2 value of the myxoid extracellular matrix. In these, ADC values of benign and malignant tumours overlap: not all malignant tumours present more cellularity than the benign and the benign often have an extracellular matrix similar to the malignant [46‐49]. This happens because conventional ADC values are calculated on a vast range of b value (b 0-600 s/mm2) and the low b values are sensitive to perfusion determined by the tumour extracellular water component as opposed to the use of high b values (300,500, 600 s/mm2) that are perfusion insensitive so perfusion effects tend to be canceled. This metric measure is called Perfusion Insensitive Diffusion Coefficent (PIDC). Costa et al. propose a PIDC value of 1.1 × 10ˉ3/s to distinguish malignant from benign lesions: malignant under this value, benign over. PIDC maps provide more accurate information about tumour tissue cellularity by minimizing vascular contributions which are higher in malignant tumours. Costa and others obtained the highest PIDC values in myxoid tumours: myxoma, myxoid liposarcoma, and low grade myxofibrosarcoma show mean PIDC values of 2.92 x 10-3 mm2/s. This value is due to the high mucin and low collagen contents of the tumour that is largely composed of water as confirmed by histological analysis [50]. When the following features are present amongst STS we can consider myxoid tumours and principally myxoid liposarcoma the main diagnostic hypothesis:

These three features are very important because myxoid liposarcomas (about 50 % of all liposarcomas) have less than 10 % mature fat, that is a low/intermediate signal intensity on T1-weighted images on conventional MR.

Small round blue cell tumours (SRBC) are a group of less differentiated and aggressive embryonal tumours with similar histologic features and immunochemistry. They include neuroblastoma, rhabdomyosarcoma, non-Hodgkin lymphoma, Ewing sarcoma. These tumours show more restricted diffusion than other STS. For this reason the differential diagnosis of a tumour with restricted diffusion on ADC maps and very low PIDC values, small round cell tumours should be the main diagnostic hypothesis, when conventional MRI and CT both with and without contrast give clues to this hypothesis.

Fibroblastic, myofibroblastic and fibrohistiocytic tumours, the most common in all age groups, usually present low signal intensity in all sequences and a flimsy to moderate signal intensity increase after contrast on conventional MRI. These signs together with increased PIDC values and a restricted diffusion on ADC maps can help in the differential diagnosis between benign and malignant masses with morphological features of fibrous tumours on conventional MRI. It must be underlined that that there are not differences in PIDC values between benign and intermediate fibrous tumours.

It is also possible to differentiate necrotic masses such as hematomas and abscesses from necrotic hemorrhagic in malignant soft tumours by ADC map values both in the central component and on the peripheral rim. Malignant tumours for their high cellularity have a more restricted diffusion on ADC maps in the peripheral rim than in the central necrotic area. This measure associated to the typical morphological features of hematomas on conventional time based T1- and T2- MR images allow to differentiate benign from malignant necrosis.

Giant Cell Tumour is a bone and soft tissue tumour presenting low PIDC values and restricted diffusion on ADC maps. These parameters can be useful in the diagnosis and management of local recurrences by revealing differentiation of post-surgical fibrosis from recurrences [51].

DWI together with conventional MR imaging and CT can help differentiate between pus-like fluids and serous cystic fluids. Serous and pure water like necrosis show active water diffusion and high ADC values, purulent pus-like fluids show restricted water diffusion and low ADC values for the high viscosity and cellularity of pus (Fig. 5a-d) (Fig. 6a-d).