The optimal use of PET/CT in the management of lymphoma patients

Abstract

¹⁸F-fluoro-deoxyglucose positron emission tomography (PET)/computed tomography (CT) scans play an important role in the management of lymphoma patients. They are critical to accurately stage disease and assess its response to therapy. In addition, PET/CT scans enable precise target delineation for radiation therapy planning. In this review, we describe the use of PET/CT scans in lymphoma, with a focus on their role in staging disease, assessing response to therapy, predicting prognosis, and planning RT.

Introduction

¹⁸F-fluoro-deoxyglucose (FDG) positron emission tomography (PET)/computed tomography (CT) scans play a central role in the management of lymphoma patients. These imaging studies are critical to accurately stage disease, assess response to therapy, and map areas to target with radiation therapy (RT). In this review, we summarize the optimal use of PET/CT for staging lymphoma, the importance of PET/CT for RT planning, the prognostic information provided by PET/CT, radiographic response criteria and their change over time, and limitations of PET/CT.

Optimal acquisition of PET/CT for staging

FDG PET is highly sensitive for lymphoma and contributes to accurate staging of disease. In series of patients with HL, DLBCL, or T-cell lymphomas, FDG PET led to upstaging in up to 5–50% of cases compared to CT alone.^1–5 As one example, in 1,171 patients treated for advanced-stage HL on a prospective trial, FDG PET resulted in upstaging of 14% of patients, primarily due to the identification of pathologic FDG uptake in the bone marrow or in anatomically normal-sized lymph nodes.¹ In this same study, PET led to downstaging in 6% of patients, primarily due to lack of pathologic FDG uptake in an enlarged spleen or lymph nodes.¹ Accurate staging is essential because therapeutic strategies and prognosis vary by disease stage. Therefore, the Lugano criteria recommend that FDG PET/CT scans be used for the initial staging of all FDG-avid lymphomas.^6,7 The vast majority of lymphomas are FDG-avid and should be staged by FDG PET/CT; possible exceptions include indolent non-Hodgkin lymphomas such as marginal zone lymphoma.

Ideally, high-quality scans are performed with intravenous (IV) contrast for optimal delineation of the sites of disease. Oral contrast may be used but it should not contain sugar that would affect FDG uptake (e.g., 250 ml Ioxitalamat solution, 12.6 mg ml⁻¹, may be given orally 45 min before the scan). In addition, because these scans are used for RT planning, it is ideal, although not always possible, for the scan to be performed on a flat table with the patient in the position that will be used for RT treatment (e.g., with arms up), as described below. Exact scan parameters depend upon the model of the PET/CT scanner and should be determined at each individual institution by balancing the as-low-as-reasonably-achievable (ALARA) principle with image quality (Table 1). Some vendors offer options to select optimal patient-specific kVp and automatic exposure control of mAs (e.g., CARE kV and CARE Dose from Siemens Healthineers).

Table 1.

Example parameters for the CT component of PET/CT for staging scans. To reduce the dose and standardize image quality, the mAs is adjusted throughout the scan to achieve a level of image noise relative to the quality reference mAs, and the kVp is selected for the scan (from a limited set of options) to optimize image contrast (CARE Dose and CARE kV, Siemens Healthineers). Actual parameters should be institution-specific

CT Parameter	Value
Quality reference mAs	200
Quality reference kVp	120
Collimation (mm)	0.6 × 64
Reconstructed slice thickness (mm)	2
Pitch	0.8
Rotation time (s)	0.5

Below we describe a workflow for obtaining optimal baseline PET/CT scans for patients with lymphoma. Ideally, the scans are performed with the patient in the position that will be used for RT, so radiation may be delivered using the highly conformal involved nodal RT technique (INRT).⁸ When upfront imaging has not been performed in the RT position, involved site RT (ISRT) is applied, using larger margins that account for uncertainties due to factors such as differences in body position. To obtain a PET in the RT position, patients should lie on a flat table and breast board with arms up (Figure 1). A 10–15 degree incline board may be used and is particularly useful for female patients with pendulous breasts to reduce radiation dose to the breast tissue during RT.⁹ The scans are performed in both free breathing (FB) and deep inspiration breath hold (DIBH). Patients are injected with 4 MBq/(kg body weight) of FDG 1 hour before scanning. The DIBH process is described to the patient, who is instructed to breathe to a comfortably deep level. The standard whole-body examination is performed under FB, and a limited PET/CT scan is acquired with DIBH for a single PET bed position over the mediastinal/thoracic region. The patient is provided with visual guidance displayed on a screen (RPM^®, Varian Medical Systems, Palo Alto, USA) and holds his/her breath for six repeated breath holds of 20 seconds each for the PET in DIBH and one breath hold for the CT in DIBH. Lastly, the six independent PET scans in DIBH are summed in a postprocessing step into a single PET dataset comparable to a full 120 second scan. The reconstruction protocol for the PET scan in DIBH is identical to a standard routine clinical PET (3D-OSEM with PSF and time-of-flight, two iterations 24 subsets, 2 mm Gaussian postfilter on 400 × 400 voxels).

Figure 1.

Example of a patient having a PET scan performed in the position that will be used for RT (i.e., on a flat table and breast board with arms up).

The workflow for obtaining pre-chemotherapy PET/CT scans in FB and DIBH is described, step-by-step, below (using a Siemens scanner: Biograph mCT, where the CT part corresponds to a Definition AS, 128 slice):

• The patient is asked to void his/her bladder.

• The patient is positioned with arms up on a breast board and a flat RT-type table. The patient may be positioned at an incline of 10–15 degrees; this incline is most important for females with pendulous breasts, so the breast tissue falls away for the superior mediastinal target.⁹

• CT Topogram:

  • Assess the attenuation, which is used to adjust the mAs throughout the scan.
• CT Pre-Monitoring:

  • Identify and mark the aorta on the appropriate slice.
• CT-monitoring:

  • Inject IV contrast.
  • Measure the Hounsfield units (HU) in the marked aorta.
  • At >80 HU, initiate the whole-body CT.

• Whole-Body CT (Table 1, e.g., parameters).

• Whole body PET:

  • Time per bed position depends on the patient’s BMI.
• DIBH CT (Table 1 for example parameters):

  • Scan region equal to the length of one PET bed, centred over the patient’s thorax.
  • Ask the patient to inspire to a comfortable level and hold his/her breath using visual feedback for the entire scan.
• DIBH PET:

  • Match the DIBH CT scan region with the bed position of the DIBH PET scan.
  • Ask the patient to inspire to a comfortable level and hold his/her breath using visual feedback for a PET acquisition of 20 s.
  • Repeat six times.
  • Complete postprocessing offline to combine the 6 DIBH PET acquisitions.

Importance of PET/CT for RT Planning

PET/CT findings are instrumental in RT planning. First, upfront staging PET/CT scans contribute to the selection of appropriate patients for RT. As one example, RT monotherapy may be recommended for patients with localized low-grade follicular lymphoma, localized nodular lymphocyte predominant Hodgkin lymphoma, or solitary plasmacytoma. Because these patients receive local therapy alone, the presence of distant disease must be excluded before RT is initiated. Multiple studies have demonstrated that PET/CT findings may up-stage disease, so these scans are useful for identifying patients with localized disease who are appropriate candidates for RT alone.^10–13 Second, post-chemotherapy PET/CT findings inform the radiation dose by revealing the disease’s response to chemotherapy. In both HL and NHL, lower radiation doses are used in the setting of a complete metabolic response (CMR), whereas higher doses are used in the setting of residual disease as revealed by PET/CT.^14,15 Lastly, PET/CT scans enable accurate delineation of disease for RT planning. Multiple researchers have shown that consideration of upfront PET data significantly influences RT target volumes when compared to target delineation based upon CT alone.^16–18

Highly accurate target delineation, facilitated by the initial PET/CT, is particularly important in the setting of modern, conformal INRT/ISRT. Below, we describe a workflow for simulation and target volume delineation that optimally incorporates baseline PET/CT data:

• Positioning for RT must be reproducible for the simulation scan and each treatment session. Patients with lymphoma in the thorax are positioned on a breast board with a pillow under the knees. A vacuum fixation cushion can be used if the patient cannot be positioned comfortably on the breast board.

• We recommend for patients to be simulated and treated with a DIBH technique.¹⁴ With DIBH, the lungs expand and the heart moves inferiorly. This shift in anatomy reduces the dose to these normal tissues during RT, which, in turn, is estimated to significantly reduce the risk of late toxicities.^19–25 For DIBH, the patient is asked to hold his/her breath at a comfortably deep level and is provided visual guidance (RPM^®, Varian Medical Systems, Palo Alto, USA).

• The planning CT scan may be performed with IV contrast but should not use oral contrast due to the difficulty of overriding the density to maintain an accurate dose calculation. The IV contrast is given with a scanning delay of approximately 30 seconds to allow contrast to circulate to the arteries and veins to assist with delineation of the tumour and organs at risk.

• The upfront PET/CT scan is fused with the post-chemotherapy RT planning CT scan(s). The scans can be matched ideally if they utilize the same setup position and breathing technique (i.e., staging PET/CT in DIBH fused with planning CT in DIBH).

• After fusion, the pre-chemotherapy GTV is delineated on the baseline PET/CT and transferred to the post-chemotherapy planning CT scan. The clinical target volume (CTV) is contoured on the planning CT (Figure 2). A planning target volume (PTV) expansion should be chosen with respect to each institution’s process and uncertainties. A margin of 0.5 to 1 cm may be appropriate for treatment with a DIBH technique, while a larger margin (e.g., 1 to 1.5 cm) cm may be necessary for treatment in FB.

Figure 2.

Example of imaging and contours for a patient with mediastinal lymphoma. The CT part of the pre-chemotherapy PET/CT in DIBH is shown on the left with the PET-positive region shown in cyan and the pre-chemotherapy GTV (delineated using information from both the PET and the CT) shown in red. The post-chemotherapy radiotherapy planning CT in DIBH is shown on the right. The images have been registered and the contours from the pre-chemotherapy scan have been copied to the post-chemotherapy scan for reference. The CTV used for radiotherapy planning is shown in pink.

Baseline imaging studies: Prognostic information

Findings from baseline imaging studies are associated with prognosis in lymphoma patients. For many decades, it has been known that the presence anatomically bulky disease, based on various definitions, is associated with oncologic outcomes. The Cotswolds revision of the Ann Arbor staging system defined mediastinal bulk as a mass measuring ≥1/3 the internal thoracic diameter at the level of T5/6 on a standing postero-anterior (PA) chest X-ray (CXR).²⁶ Multiple early studies showed that patients with bulky mediastinal disease based upon this definition experienced inferior outcomes.^27–29 Subsequent studies identified an association between disease relapse and baseline maximal tumour diameter ranging from >5 to 10 cm in the transverse plane in patients with both Hodgkin lymphoma (HL)^30,31 and non-Hodgkin lymphoma (NHL).^32,33 Of note, in diffuse large B-cell lymphoma (DLBCL), the adverse prognostic effect of maximal tumour diameter on event-free survival (EFS) and overall survival (OS) is linear in the range of 5–10 cm, so any cut-point in this range is likely to be associated with outcome.³³ With advances in CT reconstruction, measurements in the cranial-caudal dimension, as well as the axial dimension, became possible, and researchers identified an optimal cut-off of >7 cm, measured in either the transverse or coronal plane, to distinguish patients with HL at higher risk of relapse.³⁴ In summary, multiple studies have shown that baseline anatomic disease bulk, based on various unidimensional measurements, is associated with inferior outcome.

More recently, the routine use of PET/CT in lymphoma has facilitated the collection of quantitative parameters that combine functional and anatomical information that have been associated with prognosis. These measurements include metabolic tumour volume (MTV), defined as the total disease volume with an SUV exceeding a specific threshold, and total glycolysis (TLG), defined as the product of the MTV and the mean SUV of that MTV. Studies across a range of histologic subtypes of lymphoma have shown that these measures of metabolic tumour burden are highly prognostic.^35–43 As one example, in patients with early-stage HL treated on the EORTC H10 trial, Cottereau et al reported that a higher MTV was associated with inferior progression-free survival (PFS; p < 0.0001) and OS (p = 0.0001).³⁷ Similarly, in patients advanced-stage DLBCL treated on the LNH073B trial, a high baseline MTV was associated with inferior PFS (p = 0.027) and OS (p = 0.0007).⁴⁴

Taken together, these findings suggest that anatomical and functional data from baseline imaging studies are prognostically important. Parameters that incorporate both anatomical and functional information, such as MTV and TLG, may enable optimal risk stratification. Current research efforts aim to standardize these measurements and improve automation to increase the ease and speed of their acquisition.⁴⁵ We predict that ultimately measurements such as MTV will be incorporated into clinical practice to quantify upfront disease burden and contribute to accurate risk stratification.

Post-Treatment Imaging Studies: Response Criteria and their Evolution

PET/CT scans play an important role not only in initial staging of lymphomas but also in assessing disease response to therapy. Standardized criteria have been proposed for radiographic response assessments, and guidelines have evolved over time. First, in 1999, the International Working Group defined categories of response based on CT imaging. These included complete response (CR), CR unconfirmed (CRu), partial response (PR), stable disease (SD), relapsed disease and progressive disease.⁴⁶ Subsequently, PET became a standard imaging study for response assessment of lymphomas. In 2007, the International Harmonization Project defined PET positivity as FDG-avidity exceeding that of the mediastinal blood pool for residual masses ≥ 2 cm or that of the surrounding background for smaller residual masses.⁴⁷ Then, the Deauville criteria,⁴⁸ which were endorsed by the Lugano classification,^6,7 defined disease response using a 5-point scale:

• Deauville score 1: no residual FDG-avidity

• Deauville score 2: FDG-avidity ≤mediastinal blood pool

• Deauville score 3: FDG-avidity >mediastinal blood pool but ≤liver

• Deauville score 4: FDG-avidity moderately >liver

• Deauville score 5: FDG-avidity ≥2–3x greater than liver, or new sites of FDG-avid disease

It is recommended that the nuclear medicine radiologist assigns a Deauville score based on visual assessment, as well as confirmatory measurements of the SUV_max of the suspected region of disease involvement compared to mediastinum and liver. Based upon this 5-point scale, 4 categories of response have been defined^6,7 :

• Complete metabolic response (CMR): Deauville score of 1–3;

• Partial metabolic response: Deauville score of 4–5 with reduced FDG-avidity;

• No metabolic response: Deauville score of 4–5 without significant change in FDG-avidity;

• Progressive metabolic disease: Deauville score of 4–5 with increased FDG-avidity or new lesions.

A Deauville score of 1–3 is defined as negative and a score of 4–5 as positive for disease. Importantly, although a Deauville score of 3 is categorized as a CMR, some trials exploring de-escalation of therapy have conservatively defined it as an inadequate response to avoid under treatment.^49–52

Post-treatment Imaging Studies: Prognostic Information

Response assessments performed both at interim timepoints and after the completion of chemotherapy have been associated with patient outcomes. Findings on interim PET scans, performed after the initiation but before the completion of chemotherapy, have been associated with outcomes by multiple research groups. For example, one study explored the prognostic significance of an interim PET performed after two cycles of chemotherapy (PET-2) in patients with HL that was primarily advanced stage. Most patients were treated with six cycles of ABVD (doxorubicin, bleomycin, vinblastine, dacarbazine) and consolidation RT to sites with initial bulk or a residual mass. No treatment change was allowed based on the PET-2 results. The 2-year PFS for patients with a positive PET-2 was 12.8% and for patients with a negative PET-2 was 95.0% (p < 0.0001). PET-2 findings were more strongly associated with prognosis than other conventional risk factors.⁵³ Similar findings have been reported in aggressive NHL.^54,55

Similarly, PET findings after the completion of chemotherapy are significantly associated with prognosis in HL and NHL. For example, in a cohort of patients with non-bulky, limited-stage HL, the end-of-chemotherapy PET was predictive of outcome, with a 4-year PFS of 94% in PET-negative patients vs 54% in PET-positive patients (p < 0.0001) and 4-year OS of 100% vs 84%, respectively (p < 0.0001).⁵⁶ Concordantly, in a study of patients with advanced-stage HL treated with BEACOPP (bleomycin, etoposide, Adriamycin, cyclophosphamide, Oncovin, procarbazine, prednisone), the negative predictive value (NPV) of the end-of-chemotherapy PET scan was 94.1% at 12 months.⁵⁷ Although the NPV of PET is high, it must be noted that studies testing the omission of RT in patients with HL and a negative PET after chemotherapy have shown a small but consistent benefit to giving RT to sites of initial involvement.^49,51,52 These results suggest that some patients have residual microscopic disease despite having negative findings on post-chemotherapy PET scans.

The NPV of a post-chemotherapy PET is high in NHL, as well; however, the PPV of PET may be low, particularly in certain settings. For example, in a cohort of patients with primary mediastinal B-cell lymphoma treated with dose-adjusted R-EPOCH (rituximab, etoposide, prednisolone, Oncovin, cyclophosphamide, Hydroxydaunorubicin), the end-of-chemotherapy PET scan had a NPV of 98% but a PPV of only 20%.⁵⁸ Given the low PPV of PET in this setting, a confirmatory biopsy is highly recommended if persistent disease is suspected.

Interim FDG PET/CT Response Adapted Strategies

The association of interim PET findings with prognosis has prompted multiple studies of interim PET-guided treatment strategies in both HL and NHL. Important research questions in HL have been (1) if RT can be omitted in early-stage disease after a CMR to chemotherapy at interim evaluation, and (2) if chemotherapy should be intensified in patients with residual FDG-avid disease at interim evaluation. The EORTC H10 trial was a landmark study that addressed both of these questions.⁵¹ On this study, patients with early-stage HL were randomized to standard therapy vs an interim PET-adapted approach. All patients were assessed by PET/CT after 2 cycles of ABVD. Those on the standard arm proceeded with ABVD and RT. Those on the experimental arm received a greater number of cycles of ABVD but no RT after a negative PET-2, or they received BEACOPP chemotherapy and RT after a positive PET-2. For patients with a negative interim PET, the chemotherapy-only arm was closed early due to futility, and final analyses confirmed that omission of RT resulted in a significantly higher rate of events. Patients with a positive interim PET experienced superior PFS with BEACOPP and RT compared to standard ABVD and RT. Taken together, these findings suggest that omission of RT results in an increased risk of lymphoma relapse even in the setting of a negative PET-2, and that escalation of chemotherapy to BEACOPP results in superior disease control in the setting of a positive PET-2.⁵¹ Other studies have reported similar findings. For example, the UK RAPID and GHSG HD16 studies both failed to demonstrate non-inferiority of chemotherapy alone compared to combined modality therapy in early-stage HL, even in the setting of a negative PET scan.^49,52 Also, several studies have demonstrated favourable outcomes in advanced-stage HL patients with a positive interim PET whose therapy was escalated to BEACOPP.^59–61 Based on these findings, a PET2-guided treatment strategy is frequently used in the management of patients with HL.

While interim PET-adapted strategies have gained hold in HL, the data in NHL are less conclusive. Numerous studies have explored escalation of therapy in response to a positive interim PET in this patient population.^62–69 However, no consistent improvement in outcomes has been demonstrated with intensified regimens compared to standard R-CHOP. Therefore, in patients with DLBCL, treatment escalation for interim PET positivity is not recommended in routine practice.

Limitations of PET in lymphoma

Multiple factors may cause false radiographic findings. For example, PET scans may give false-positive findings if they are performed too soon after the completion of treatment. Therefore, the Imaging Subcommittee of the International Harmonization Project in Lymphoma recommends waiting to perform a PET scan until ≥3 weeks, and preferably ≥6–8 weeks, after chemotherapy and ≥8–12 weeks after RT.⁴⁷ In addition, a few conditions mimic lymphoma in PET imaging. Therefore, the nuclear medicine radiologist should not use a fixed SUV cut-off value to define disease, but instead should rely on visual analysis and clinical understanding of disease in reading and interpretation. One cause of false-positive PET findings is physiologic uptake in brown fat, especially in young patients. CT with IV contrast is useful to distinguish the physiological uptake in brown fat from lymphoma-involved nodes. In addition, medications such as propranolol or fentanyl can reduce brown fat uptake of FDG. False-positive findings may occur due to physiologic uptake in other metabolically active tissues, as well, such as Waldeyer’s ring or the salivary glands. FDG-avidity may be observed in the setting of sarcoidosis, thymic rebound and bone marrow activation. In addition, prior sites of skeletal involvement by lymphoma may show persistent FDG-avidity after treatment due to bone remodelling. Lastly, FDG-activity may indicate a concurrent malignant tumour. Therefore, it is essential that relevant clinical history and previous imaging is provided when interpreting PET scans for lymphoma.

False-negative findings are possible, as well. For example, due to the limited spatial resolution of PET imaging, partial volume effects may lead to false-negatives for small lymph nodes (<5–6 mm). Furthermore, lymphoma with low-grade metabolic activity, such as MALT lymphoma, may be difficult to detect with PET.

Conclusions

PET/CT scans play a crucial role in the management of lymphoma patients. Over the past decades, cure rates have soared and toxicity rates have declined. PET/CT scans have been integral to these improvements by contributing to rigorous staging, appropriate selection of therapy, accurate response assessment, and precise target delineation for RT planning.