Monitoring Therapy in Bone and Soft Tissue Tumors

Pathophysiology

The histologic response to neoadjuvant chemotherapy is one of the most important prognostic factors for patients with bone or soft tissue sarcoma who are scheduled for surgical treatment. Patients whose tumors show little necrosis relative to the fraction of viable tumor after neoadjuvant chemotherapy have poorer survival than patients with tumors that have more chemotherapy-induced necrosis. The amount of spontaneous, or non–chemotherapy-induced, necrosis as a result of the tumor outgrowing its blood supply can be quite substantial. Therefore, both histologic and radiologic techniques focus on determination of the fraction of the entire tumor that is still viable instead of the fraction that is necrotic. Historically, only one or two macrosections of the resected tumor have been used in the pathologic laboratory to determine the fraction of viable tumor. To avoid sample errors that are secondary to this limited analysis, MRI can presently be used by the pathologist to target components of the tumor that are viable, thereby avoiding these errors.

Although the principle of classifying response of musculoskeletal sarcoma to therapy, based on fraction of viable tumor, is the same for all sarcomas, some differences do exist. The basis for the classification system has initially been described for osteosarcoma ( eTable 100-1 ). Residual viable tumor in osteosarcoma preferentially persists within the soft tissues, cortical bone and endosteal surface, zones adjacent to cartilage, ligaments, and areas around zones of liquefaction.

eTABLE 100-1

Histologic Classification of Response to Chemotherapy in Osteosarcoma

Data from Huvos AG, Rosen G, Marcove RC. Primary osteogenic sarcoma: pathologic aspects in 20 patients after treatment with chemotherapy en bloc resection, and prosthetic bone replacement. Arch Pathol Lab Med 1977;101:14-18.

Grade	Necrosis (%)	Histologic Appearance
I	0-49	Little or no necrosis
II	50-89	Areas of acellular tumor osteoid and/or fibrotic material attributable to the effect of chemotherapy, with other areas of viable tumor
III	90-99	Predominant areas of acellular tumor osteoid and/or fibrotic material attributable to the effect of chemotherapy, with only scattered foci of viable tumor cells
IV	100	No pathologic evidence of viable tumor within the specimen

The grading system is slightly modified for Ewing sarcoma ( eTable 100-2 ). Sites of predilection for persistence of minimal residual viable tumor in Ewing sarcoma comprise the subperiosteal area, the intramedullary compartment, the soft tissues, and the zones adjacent to hemorrhagic areas. The Bologna group reports on a slightly different grading system in patients with Ewing sarcoma. For the evaluation of chemotherapy-induced necrosis, they use a three-grade scale, in which grade I corresponds to a tumor with at least one macroscopic nodule of viable tumor tissue, grade II corresponds to a tumor with only isolated microscopic foci of viable tumor tissue (in total, not more than one field at 10× magnification), and grade III corresponds to a tumor in which no viable tumor tissue can be found. The probability of developing recurrent tumors is slightly greater than 50% for high-risk patients (grades I and II) compared with 5% for low-risk patients (grade III).

eTABLE 100-2

Histologic Classification of Response to Chemotherapy in Ewing Sarcoma of Bone

Data from van der Woude HJ, Bloem JL, van Oostayen JA, et al. Treatment of high-grade bone sarcomas with neoadjuvant chemotherapy: the utility of sequential color Doppler sonography in predicting histopathologic response. AJR Am J Roentgenol 1995;165:125-33.

Poor Response
Class I	Minimal or no effect of chemotherapy: ≥90% viable tumor, <10% tumor necrosis
Class II	Moderate effect: solid areas of viable tumor remnants, 10%-90% tumor necrosis or loose hypocellular fibrous tissue
Good Response
Class IIIa	Minimal residual disease: <10% viable tumor, localized in small clusters subperiosteally between trabeculae of reactive periosteal bone formation and/or in the soft tissues
Class IIIb	Minimal residual disease: <10% viable tumor, localized in small clusters in the intramedullary tumor compartment only
Class IV	No viable tumor cells, only necrosis and/or vascularized fibrous tissue

Although analysis of resected soft tissue tumors is hampered by the absence of osseous support and structure of the specimen, a similar grading system has been used for soft tissue sarcoma.

Recurrence of tumor is typically diagnosed when residual tumor is detected by imaging techniques or when it becomes symptomatic as a result of growth. Recurrences of both osseous and soft tissue sarcoma typically occur in the soft tissues of the surgical bed. Recurrences of benign osseous tumors are frequently also found in bone because resection margins are smaller than in sarcoma surgery.

Imaging Techniques

Dynamic Contrast-Enhanced Magnetic Resonance Imaging

Dynamic contrast-enhanced (DCE) MRI has evolved into an adjunct imaging technique that can be integrated into a standard morphologic imaging protocol. The small molecular contrast agent with gadolinium passes from the intravascular space into the interstitium at a rate determined by the permeability of the capillaries, total vascular cross-sectional area, interstitial pressure, volume of extracellular space, contrast agent injection rate, and cardiac output. In addition, there are several technical MR parameters that influence the measured signal intensities, posing a challenge for quantitative comparison of data generated by different MR systems. In DCE MRI, rapid imaging techniques are used to acquire serial images before, during, and after a small, intravenous bolus injection of Gd-chelates, such as Gd-DTPA, to evaluate the initial distribution of the contrast agent in the capillaries and interstitial space ( eFig. 100-1 ). Depending on concentration, Gd-chelates decrease T1 and T2 relaxation times. At high concentrations, T2 shortening is dominant, whereas at lower concentrations, T1 shortening is dominant. Because the concentration of Gd-chelate is initially high, first-pass T2-weighted imaging is sensitive to blood flow and blood volume. T1-weighted imaging is sensitive to the low concentrations of Gd-chelates that permeate through the capillary walls and is thus used to measure parameters that reflect permeability. Directly after MR data acquisition is started, a bolus injection of Gd-chelate with a concentration of 0.2 mL/kg at an injection rate of 4 mL/s is intravenously administered. An imaging frequency (temporal resolution) of at least one image per 3 seconds is initially needed to produce at least two data points in the first part of the enhancement curve, with its potentially rapid rise of concentration of the contrast agent in the tumor. The entire first phase lasts 7 to 15 minutes. In addition to obtaining information on the initial enhancement, information on distribution of the contrast agent is obtained by sampling for approximately 5 minutes with a temporal resolution that can be lower than the initial one. Adequate tumor sampling with multiple slices, high spatial resolution, and high temporal resolution are competing parameters. Higher temporal resolution appears to improve the specificity of examinations because of better delineation of differences of signal intensity on the time-intensity curves. High spatial resolution imaging with sufficient high temporal resolution and adequate tumor coverage may improve the detection of small foci of residual viable tumor. Fast or ultrafast MR sequences, such as turbo field-echo (Philips, Best, The Netherlands), turbo fast low-angle shot (FLASH) (Siemens, Erlangen, Germany), and inversion recovery prepared fast gradient-recalled acquisition in the steady state (GRASS) (General Electric, Milwaukee, Wisc.) allow fast imaging with a sufficiently high temporal and spatial resolution.

eFIGURE 100-1, Factors determining early tissue enhancement. The lower parts of A to D show what occurs at the level of the capillary and the interstitial space after intravenous bolus injection; the upper parts graphically display the changes in signal intensity (SI) interval between the intravenous bolus injection and arrival of the bolus in the capillary, which is determined by the injection rate, the heart rate, the localization of the lesion, and the local capillary resistance (tissue perfusion). C , The enhancement rate during the first pass of the contrast agent is determined by number of vessels (tissue vascularization), local capillary resistance (tissue perfusion), and capillary permeability. Tissues with high vascularization, perfusion, and capillary permeability (X) will enhance earlier and faster than tissues with fewer vessels, higher capillary resistance, and lower capillary permeability (Y) . C , After the first pass of the bolus, the signal intensity increases further until the concentration of the gadolinium contrast medium in the blood and the interstitial space of the tissue is equal. In tissues with a small (S) interstitial space, this equilibrium is reached earlier than in tissues with a larger (L) interstitial space. D , As the arterial concentration of the contrast medium decreases, the SI drops while the gadolinium is progressively washed out from the interstitial space. This process occurs faster in tissues with a small (S) interstitial space than in tissues with a large (L) interstitial space.

Enhancement on T1-weighted DCE MRI can be assessed in two ways: by the analysis of signal intensity changes (semiquantitative) and/or by quantifying contrast agent concentration change using pharmacokinetic modeling techniques. Signal enhancement can be visually evaluated with the use of subtraction techniques and by using the ROI method. Subtraction images are created by subtraction of precontrast MR images from the Gd-chelate–enhanced MR images. On these subtraction images, small areas within the tumor that enhance fast relative to enhancement of nearby arteries can be identified. ROIs are placed on these early enhancing areas that are identified on subtraction images. The signal intensity within these ROIs is plotted against time in TICs ( eFig. 100-2 ). Semiquantitative parameters can be derived from these TICs. These parameters include curve shape, onset time (time from arrival of contrast agent in an artery relative to arrival of contrast agent in the tissue of interest), slope of enhancement curves, time to maximum enhancement, maximum enhancement, and wash-out ( eFig. 100-3 ). Commercial software is becoming available that generates standardized quantitative data. This requires T1 mapping before contrast agent administration.

eFIGURE 100-2, Classification for subjective assessment of time-intensity curves: type I, no enhancement; type II, gradual increase of enhancement; type III, rapid initial enhancement followed by a plateau phase; type IV, rapid initial enhancement followed by a wash-out phase; and type V, rapid initial enhancement, followed by sustained late enhancement.

eFIGURE 100-3, Time-intensity curve (TIC). In a TIC, the temporal change of the signal intensity in a region of interest (ROI; or pixel) is plotted against time. At T start when the bolus enters the ROI, the signal intensity rises above the baseline signal intensity (SI base ). The steepest slope represents the highest enhancement rate during the first pass (wash-in rate) and is mainly determined by tissue vascularization, perfusion, and capillary permeability. At T max , the time of maximum enhancement, capillary and interstitial concentrations reach equilibrium. The time period between the end of the first pass and the maximum enhancement is mainly determined by the volume of the interstitial space. The wash-out rate can be calculated from the negative slope of the curve. a.u., Arbitrary units; T, time interval between SI end and SI prior .

These quantitative techniques using pharmacokinetic modeling are preferred. Signal intensity values observed during dynamic acquisition can be used to estimate contrast agent concentration at each time point. Mathematically fitting these data to pharmacokinetic models yields quantitative kinetic parameters. Pharmacokinetic modeling of data in each voxel has been used to quantify the hemodynamic parameters. In first-pass or slope images, all pixels are displayed with intensity equal to the highest enhancement rate (i.e., during the first pass).

A two-compartment model relating the change of tissue tracer concentration to the difference between arterial plasma and interstitial fluid concentrations is most often used. A detailed discussion on pharmacokinetic modeling techniques is beyond the scope of this chapter and can be found elsewhere. A transfer constant K _trans describes the transendothelial transport of the small molecular contrast agent from the vascular to the interstitial space. Three major factors determine the behavior of the contrast agent in tissues during the first few minutes after injection: blood perfusion, transport of contrast agent across vessel walls, and diffusion of contrast medium in the interstitial space. Changes in the semiquantitative parameters can be used to assess changes in microcirculation during chemotherapy.

Diffusion-Weighted Magnetic Resonance Imaging

In addition to information based on intrinsic MR contrast mechanisms and contrast-enhanced MRI, diffusion-weighted MRI is able to provide an estimate of molecular diffusion in tissues. Diffusion-weighted MRI uses a pulse sequence that is sensitive to motion of protons, also called brownian motion. The extracellular, intracellular, and transcellular random motion of water molecules as well as microcirculation (perfusion) contribute to the MR signal in diffusion MRI. The extracellular component and perfusion contribute most to the signal on diffusion-weighted MRI. However, diffusion-weighted sequences are sensitive not only to molecular motion but also to bulk motion (moving patient), pulsation of vessels, and intrinsic contrast mechanisms such as T2 contrast (the so-called T2 shine-through). These contrast mechanisms and artifacts can be reduced by special measures in pulse sequence design, such as electrocardiographic triggering to reduce pulsation artifacts, fast scanning to reduce bulk motion artifacts, and short echo time (TE) to reduce T2 shine-through, but they also must be taken into account when analyzing diffusion-weighted images. Different diffusion-weighted sequences have been described: spin-echo, echo-planar imaging, and steady-state free precession (SSFP) sequences.

Only limited reports have been published describing the therapeutic response of osteosarcoma by diffusion-weighted MR imaging. Postchemotherapy changes, such as cell membrane injury, cell death, or a reduction in cell density, cause an expansion of the extracellular diffusion space and lead to a greater degree of unrestricted extracellular water motion. Moreover, there will be more diffusion in necrotic tissue than in viable tumor because there are fewer cell boundaries and a larger extracellular space. More diffusion means more destructive dephasing of spins, resulting in lower signal intensity. Increased diffusion in necrotic tissue can be displayed as signal loss on diffusion images relative to images without the gradient settings that introduce sensitivity to diffusion. Alternatively, an increase in diffusion can be displayed as an increase in signal on apparent diffusion coefficient (ADC) images.

Ultrasonography

Color Doppler ultrasonography can provide clinically relevant information related to tumor vasculature but is currently not widely used for monitoring response to chemotherapy in musculoskeletal tumors. Ultrasonography is a low-cost, noninvasive, and readily available modality, but considerable operator expertise is required, and the technique is prone to interobserver and intraobserver variation. Doppler ultrasonography can depict blood flow only in relatively large vessels and is not sensitive to blood vessels smaller than 100 µm in diameter. Absence of detectable intratumoral blood flow thus not only can reflect absence of flow but also may be secondary to flow that is below the minimal threshold of detection.

In malignant tumors, two different Doppler signals have been identified that may coexist in the same tumor. The first is a high-systolic Doppler shift corresponding to high peak-systolic velocities with or without enhanced diastolic flow arising from arteriovenous shunting. The second Doppler signal is a low-impedance signal with little or no variation between systole and diastole and spectral broadening, which is associated with low-resistance flow secondary to the presence of thin-walled sinusoidal spaces. Angiogenesis results in the formation of tumor vessels that lack a muscular wall and that are therefore incapable of building up hemodynamic resistance. Both Doppler signals thus relate to the histologic structure and hemodynamic properties of tumor circulation. Abnormal newly formed tumor vessels are most prevalent at the tumor periphery but can also be encountered in the center of the tumor.

Recent technical improvements of instrumentation with higher signal-to-noise ratio, as well as the advent of micron-sized gas-filled microbubble-based contrast agents that are true intravascular contrast agents, will improve the threshold to lower blood volume and slower flow and give direct and indirect information on tumor neovascularity and response to therapy. The results of estimating blood volume by ultrasound imaging using microbubble contrast agents have been shown to correlate with estimates derived from MR data. Visualization of the vasculature may further be enhanced with the use of harmonic imaging, which uses the ability of microbubbles to oscillate nonlinearly. In addition, microbubbles labeled with agents that bind to angiogenic markers, such as α _v β ₃ integrin, are useful for molecular imaging of angiogenic vasculature. Animal studies and phase I clinical trials have demonstrated differences in flow and vascular volume in tumors at different stages during antiangiogenic therapy. Prospective clinical studies are needed to determine whether this will result in real diagnostic improvement.