Evaluation of baseline lesions
Positron Emission Tomography (PET), particularly with 18-Fluorodeoxyglucose (FDG), continues to define and expand its role in oncologic management. Beyond tumor size, as definable by computed tomography (CT), PET provides a measure of metabolic activity in tumors and is integral in initial workup for multiple disease sites including head/neck squamous cell carcinoma, non-small cell lung cancer (NSCLC), lymphoma, and many others. For head and neck cancers, FDG PET imaging facilitates early detection of persistent and recurrent head/neck squamous cell carcinoma after chemoradiotherapy, increasing deferral of surgical neck dissection to the salvage setting in many cases. In the setting of non-small-cell lung cancer, PET is further considered standard of care for radiotherapy treatment planning. Post-treatment PET has further shown to facilitate assessment of treatment response, with metabolic response seen on PET pre-dating CT-based radiographic response. Though routine post-therapy PET after definitive non-surgical management is standard management for head/neck squamous cell carcinomas, evidence to support this routine use for other subsites is lacking and thus currently not recommended for various organ sites including lung. This chapter herein discusses various PET imaging techniques and assessment variables that have been used to investigate assessment of response to oncologic treatment. In particular, assessment of response with early and late post-radiotherapy PET imaging for head and neck, NSCLC, rectal cancer, esophageal cancer, and lymphoma are discussed. Recent research involving on-treatment PET imaging as well as future work are further presented.
2. PET technique
PET is a medical imaging technique employing the unique parameters of decay of positron-emitting isotopes. Today, PET is routinely used in conjunction with computed tomography (CT) in a combined medical imaging device, PET-CT, allowing anatomic image correlation with the functional imaging obtained by PET.
A number of PET radiotacers have been used in oncology, though 18F-Flourodeoxyglucose (FDG) is FDA-approved and most commonly employed. Other agents including 18F-FMISO (18F-Fluoromisonidazole), 18FLT (18F-Fluorothymidine), 16b-18F-Fluoro-5a-Dihydrotestosterone (18F-FDHT), 60Cu-ATSM (Copper-diacetyl-bis(N4-methylthiosemicarbazone)), 18F-FES (16a-18F-fluoro-17b-estradiol), 11C-MET (11C-methionine), show significant potential to monitor the response to therapy before, during, or after therapeutic intervention.
18F-FDG chemically is 2-deoxy-2-18F-fluoro-D-glucose, a glucose analog. On 18FDG, the positron-emitting radioactive isotope fluorine-18 is substituted at the 2' position of the glucose molecule preventing glycolysis, which requires a hydroxyl group at the 2’ position. It has significantly increased uptake in tissues with increased metabolic activity, in particular, most malignancies . With increased demand for gluose, tumors tend to have increased expression of glucose transport proteins at the cellular membrane as well as increased hexokinase . With its relatively short half-life of 110 minutes, in tissues with rapid uptake, the 18F decay occurs primarily when trapped intracellularly, helping visualize these areas on PET. Malignancies with moderate to high 18F-FDG uptake include most lung cancers, colorectal cancers, esophageal cancers, gastric cancers, head and neck cancers, cervical cancers, ovarian cancers, breast cancers, lymphomas, and melanoma . Hepatocellucar carcinoma, testicular cancers, renal cancers, sarcomas, and neuroendocrine tumors have variable 18F-FDG uptake . Prostate adenocarcinoma, the most common cancer in males, has generally low metabolic activity, rendering 18F-FDG particularly less helpful for this malignancy in the primary setting, leading to potential false negative interpretation [5–7]. As 18F-FDG undergoes physiologic excretion through the bladder hinders evaluation of both bladder and prostate malignancies. Overall, 18F-FDG has been the most used oncologic tracer, but its applicability is not universal across all malignancies, nor is its uptake specific to only neoplasm. Though aberrant tumor growth in malignancy routinely results in increased 18F-FDG avidity, it is not tumor specific other benign tissue and benign conditions can also have variable uptake of 18F-FDG (e.g. inflammation or hyperplastic bone marrow) potentially leading to false positive findings [4,7]. As bone marrow hyperplasia and inflammation are not uncommon consequences after oncologic treatment including surgery, radiation therapy, and/or chemotherapy, 18F-FDG PET has limitations particulary in post-therapeutic assessment.
2.2. Other radiotracers
Beyond 18F-FDG, other markers exploit other cellular mechanisms for biologic imaging with PET. Other markers have been used to assess tumor proliferation with markers of DNA synthesis. As thymidine is unique to DNA, this has been exploited with various radiotracers including 11C-thymidine—which is limited by the short half-life of 11C—as well as thymidine analogs 18F-FLT and 8F-FMAU with the longer half-life of 18F . 18F-FLT acts as a substrate of cytosolic thymidine kinase 1 (TK1), a key enzyme for salvage DNA synthesis, and 8F-FMAU is a substrate of thymidine kinase 2 (TK2), located in mitochondria, resulting in different distributions of these markers in tissue [9,10]. Although tumors tend to be less avid of 18F-FLT in comparison go 18F-FDG, tumor delineation from background tissue can be superior with 18F-FLT in regions such as the brain, mediastinum, and intestines, where normal physiologic uptake of 18F-FLT in these areas are much lower, yielding a high tumor-to-background ratio [1,11–13]. In a head-to-head comparison of 18F-FLT to 18F-FDG to assess chemotherapy response in patients with breast cancer who had imaging with both radiotracers, change in FLT uptake after one cycle of chemotherapy better predicted late changes in tumor marker levels and correlated well with eventual radiographic tumor response . Though less employed in comparison to 18F-FLT, 18F-FMAU has shown ability to visualize breast, brain, lung, and prostate tumors. As 18F-FMAU shows low uptake in normal bone marrow—as opposed to 18F-FLT, which has high bone marrow uptake—18F-FMAU is more suitable for visualization of metastatic prostate cancer.
Radiolabeled Cu-ATSM (60/62/64Cu-ATSM) and 18F-FMISO are currently the two primary radiotracers employed for imaging tissue hypoxia—correlated with decreased sensitivity to treatment—and has been with worse clinical outcomes [15,16]. 60Cu-ATSM has been found to predict aresponse to therapy for NSCLC and predict both recurrence and survival outcomes for cervical and rectal cancers [17–19]. Clinically, pretreatment 18F-FMISO has been shown to predict survival in patients with head and neck cancer and glioblastoma multiforme [20,21].
Various amino acid radiotracers have been used, with 11C-MET (a methionine analog) the most common. It has found a niche in CNS malignancies. In malignant gliomas, decreased uptake during temozolomide therapy has shown improved time to progression; areas of uptake have shown areas at high risk of recurrence, and has helped distinguish post-radiation necrosis versus recurrent malignancy [22–24].
An additional class of radiotracers have aimed to assess hormone receptors, as receptors play an integral role in malignancies, paticulary prostate and breast cancers. 18F-FES is the most commonly used, showing correlation with estrogen receptor (ER) levels as well as response to aromatase inhibitors [25,26]. Ultimately, pretreatment uptake values have shown to predict patients who would or would not respond to therapy . For prostate cancer, 18F-FDHT is an analog of 5α-dihydrotestosterone. Correlation with treatment response has not as well been shown in prostate cancer with this marker, though 18F-FDHT uptake has been associated with high PSA levels .
Historically, PET imaging was obtained with a single static set of images obtained up to 1 hour after injection of 18F-FDG. As noted previously, a diagnostic limitation of PET imaging for oncologic diagnosis are the false positive findings secondary to inflammation quite commonly associated to therapeutic response. As 18F-FDG uptake and retention kinetics are potentially different between tumor and normal tissue inflammation, people have investigated more dynamic methods of acquiring metabolic PET data.
In a series of 21 patients with head and neck carcinomas, dual-time-point 18F-FDG PET studies helped differentiate malignancy from inflammation . Standard uptake values (SUVs) of tumors were shown to increase on the second (delayed) study by mean of 12% in comparison to matched contralateral normal tissue which showed a mean decrease of 5% on delayed imaging (p<0.05) . Inflammatory sites showed relatively stable uptake over the two scans; time interval between scans correlate with tumor SUV increase; and interval of greater than 30 minutes was recommended for separation .
For evaluation of pulmonary nodules, an early study of 36 patients siwht 38 pulmonary nodules, malignant or benign, underwent dual-time-point PET at 70 and 123 minutes post-injection . A similar trend was seen with mean increase of tumor SUV of 20% (from 3.7 to 4.4) in malignant lesions from early to delayed scan (P<0.01); benign lesions showed stable and lower mean SUVs (1.1 on both early and delayed imaging) . They determine a threshold of 10% increase from early to delayed imaging as the best predictor, reaching sensitivity of 100% and specificity of 89% . Other data have shown similar trends of increased 18F-FDG uptake from first to second scan in malignant tissue and stable to decreased uptake in benign lesions .
In a study of 47 patients with suspected pancreatic cancer, patients had dual-time-point 18F-FDG PET imaging acquired 1 and 2 hours after injection; further, some patients had a third scan at the 3-hour time point after injection . Twenty-two lesions were malignant, whereas 20 were benign. With a constant SUV threshold, the initial 1-hour PET was found to be 95% sensitive, missing one of 22 malignant lesions, and 83% accurate. With addition information of 2-hour PET imaging, retention characteristics of 18F-FDG increased diagnostic accuracy to 91.5%, with no decrease in false negatives . The additional information provided by a 3-hour PET did not improve diagnostic accuracy beyond the dual-phase imaging obtained at the 1-hour and 2-hour time points .
With these potential diagnostic advantages from dual-phase PET-CT (with 2 PET scans separated by a time interval) has grown increasingly common. With the extra information provided with dual-phase imaging, people have further investigated ‘dynamic PET’ imaging, obtaining continuous PET data over time rather than at discrete or brief time spans, adding further breadth of data to kinetic profiles of uptake. Early work used dynamic continuous imaging to model discrete time-point imaging, showing linear change over time in patients with breast cancer. A recent study utilized dynamic PET imaging with 18F-FCho (18F-labelled fluoromethylcholine) to assess time-activity curves of space occupying brain lesions . Another recent study used a dynamic PET-CT approach to assess cervical adenopathy in patients with oral/head and neck cancer; consecutive imaging at nine time points with PET/CT were obtained from 60-115 minutes after injection . At our institution, we have recently initiated an adaptive radiation therapy protocol for patients with head/neck cancer in which patients receive weekly dynamic PET imaging over approximately 90 minutes during the course of treatment.
Though PET imaging acquires three-dimensional (3D) data, as CT technology has advanced to enable four-dimensional (4D) imaging with full 3D CT image sets corresponding to various portions of a respiratory cycle, so now have 4D-PET-CTs come into clinical use, with potential to reduce image smearing, improve accuracy of PET-CT co-registration, and increase the measured SUV [34,35]. A study evaluating 57 pulmonary lesions showed particular benefit in characterizing smaller tumors, with 4D studies showing higher differences in SUVmax percent difference in comparison to 3D studies (p<0.05) assessment of smaller lesion lung lesions, with better characterization . A recent study illustrated utility of respiratory-correlated 4D-PET-CT for target delineation of squamous cell carcinoma of the esophagus, further indicating SUV threshold of 20% or 2.5 for autocontouring the gross tumor volume (GTV) . Algorithms for semiautomatic contouring have also been proposed for pulmonary lesions with minimal difference (0.1 ± 0.1 mm) on phantom studies and 0.8 ± 0.2 mm on patient tumors . Four-dimensional PET/CT has been reported to facilitate planning stereotactic radiotherapy of liver metastases  and pulmonary tumors .
3. PET parameters
From an oncologic standpoint, PET imaging is notably quite useful in its ability to quantitate parameters associated with PET uptake. An assortment of quantitative values can be obtained from each scan and from multiple-time-point scans, as well as across different scans obtained at different time points with respect to treatment (e.g. pre-treatment versus post-treatment), providing valuable information for treating physicians.
A common measurement of PET images for clinicians is the semi-quantitative value referred to as “standardized uptake value (SUV) .” Standardized uptake values are calculated throughout the three-dimensional array of CT regions, with variable SUVs throughout an image. SUV provides an index of regional tracer uptake and is a function of local radioactivity concentration, injected activity, and patient’s weight. 18F-FDG SUV can help differentiate tumor from tissue, and when used, corrections to calculation are recommended . A common method of correction accounts for a patient’s lean body mass “SUVlbm,” commonly written as “SUVlbw” (lbw=“lean body weight”), “SUVlean,” or “SUL.”
Within a region of interest (ROI) on a PET-CT, various PET quantitative factors can readily be obtained. The most commonly reported value from PET-CT oncologic imaging the maximum SUV value (SUVmax). SUVmax values are measured and reported at areas concerning for malignancy (e.g. a primary tumor and associated regional lymph nodes and distant metastases as well as other highly avid areas that may represent inflammation or reactive changes). Pre-treatment SUVmax with 18F-FDG has been reported to be prognostic for many organ sites including lung [44–46], head and neck , esophagus [48,49], gastroesophageal junction  gastric , pancreas  cervix , rectum [53,54], lymphoma , and soft tissue sarcoma .
Beyond SUVmax of an ROI, the arithmetic mean SUV (SUVmean) of voxels within the ROI have been used for oncologic assessment [57–59]. New parameters, which show promise in oncologic assessment, include the metabolic tumor volume (MTV) and total glycolytic activity (TGA) [60–63]. The MTV is defined as the tumor volume based on PET uptake and can be particularly helpful in comparison to CT-imaging when background density is similar to tumor density on CT. The boundary of MTV can be defined manually or with various parameters such as a fixed SUV threshold, percentage of SUVmax (e.g. 38%, 50%, and 60%), and gradient. On pre-treatment imaging prior to radiotherapy the volume delineated by PET-fusion to planning CT effectively corresponds to the MTV, which is utilized for biologically-targeted radiotherapy [64–66]. Such methods have been used extensively for lung radiotherapy planning, where PET staging is recommended . MTV has shown to predict overall survival in lung cancer , head and neck cancer , and esophageal cancer .
Total glycolytic activity (TGA), defined as the (MTV) x (SUVmean), is the primary PET parameter that includes both both anatomic (size) as well as metabolic parameters (e.g. with 18F-FDG). In an analysis of TGA and MTV in 45 patients with oral or oropharyngeal SCC, stage, on univariate cox regression, MTV and TGA were the most associated with progression-free survival (PFS) and overall survival (OS) (p=0.002 and p=0.006, respectively), moreso than tumor grade (p=004) and SUVmax (p=0.56) .
Retention index (RI) is a dynamic parameter that can be calculated with dual-time-point (early and delayed) PET imaging, where RI is the difference of SUVmax on two scans divided by initial SUVmax. Rate of decline of RI during lung irradiation has shown to predict locoregional recurrence . Further, in an analysis of 68 women with breast cancer, in comparison to other parameters including early and delayed SUVmax, RI showed best relation to biologic parameters including grade, Ki-67, and c-erbB-2 expression .
From an oncologic standpoint, beyond the importance of baseline PET imaging for staging and radiotherapy planning, subsequent PET scans, whether during treatment or subsequent, are used for assessment of treatment response. From such data, inter-PET analysis can be performed (e.g. comparison of a pre-treatment scan to a post-treatment scan), not to be confused with factors such as the RI which are measured across two different scans performed during two time points of a single PET (e.g. early and delayed scans). Inter-PET parameters include the difference or change in (delta, Δ) values of parameters already previously discussed as well as “percent of” (e.g. percent of baseline), percent reduction from baseline, and rate of change (velocity “VEL”). Examples of such variables comparing a new PET to a baseline PET are as indicated below, where
4. Response criteria
Various methods for assessing and categorizing response of tumors based on radiographic imaging have been proposed, including the World Health Organization (WHO) criteria, the Response Evaluation Criteria in Solid Tumors (RECIST), and RECIST 1.1 [72–75]. Such criterica, depend on radiographic imaging, which may not best assess the biologic response, particularly given that metabolic response on PET routinely anatomic radiographic response on CT . Accordingly, methods of categorizing response with PET have been developed, namely the European Organization for Research and Treatment of Cancer (EORTC) criteria and newer PET Response Criteria in Solid Tumors (PERCIST, version 1.0) [43,77]. A separate metric of response definitions using 18F-FDG PET has been developed for lymphoma response and used for clinical trials . Definitions of criteria are delineated in Table 1, Table 2, Table 3, Table 4, and Table 5.
|Measurable lesions have minimum size of 10 mm by CT scan,10 mm caliper measurement by clinical exam (lesions which cannot be accurately measured with calipers should be recorded as non-measurable), or 20 mm by chest X-ray. All other lesions are considered non-measurable||Tumor regions defined on pretreatment scan should be drawn on region of high 18F-FDG uptake representing viable tumor. Whole tumor uptake should also be recorded.|
Uptake measurements should be made for mean and maximal tumor ROI counts per pixel per second calibrated as MBq/L.
Partial volume may affect measurement of 18F-FDG uptake.
Tumor size from anatomic imaging in relation to PET scanner resolution should be documented where possible.
|Measurable target lesion is hottest single tumor lesion SUVlbw of ‘‘maximal 1.2-cm diameter volume ROI in tumor” (Peak SUVlbw). Peak SUVlbw is at least 1.5-fold greater than liver SUVlbw mean +2 SDs (in 3-cm spherical ROI in normal right lobe of liver). If liver is abnormal, primary tumor should have uptake > 2.0 SUVlbw mean of blood pool in 1-cm-diameter ROI in descending thoracic aorta extended over 2-cm z-axis.|
Uptake measurements should be made for peak and maximal single-voxel tumor SUVlbw. Other SUV metrics, including SUVlbw mean at 50% or 70% of Peak SUV, can be collected as exploratory data; TLG can be collected ideally on basis of voxels more intense than 2 SDs above liver mean SUL
|Disappearance of all target lesions. Any pathological lymph nodes must have reduction in short axis to <10 mm.||Disappearance of all non-target lesions and normalization of tumor marker level. All lymph nodes must be non-pathological in size (<10 mm short axis).|
|≥ 30% decrease in the sum of diameters of target lesions, taking as reference the baseline sum diameters.||N/A|
|Neither sufficient shrinkage to qualify for PR nor sufficient increase to qualify for PD, taking as reference the smallest sum diameters while on study.||N/A|
|≥ 20% increase in the sum of diameters of target lesions, taking as reference the smallest sum on study (including the baseline sum if that is the smallest on study). In addition to the relative increase of 20%, the sum must also demonstrate an absolute increase of at least 5 mm.|
The appearance of one or more new lesions is also considered progression.
|Unequivocal progression of existing non-target lesions.|
The appearance of one or more new lesions is considered progression.
|N/A||Persistence of one or more non-target lesion(s) and/or maintenance of tumor marker level above the normal limits.|
|CR||- no detectable clinical or radiographic evidence of disease|
- no disease-related symptoms
- no biochemical abnormalities
- negative BMB (if positive before treatment)
- lymph nodes >1.5 cm at baseline regress to ≤ 1.5 cm
- lymph nodes 1.1-1.5 cm at baseline regress to ≤ 1.0 cm
|-CR by IWC with a completely negative PET|
- CRu, PR, or SD by IWC with a completely negative PET and negative BMB if positive prior to therapy
- PD by IWC with a completely negative PET and CT abnormalities (new lesion or increasing size of previous lesion) ≥ 1.5 cm (≥ 1.0 cm in the lungs) and negative BMB if positive prior to therapy
|CRu||- same as CR but either residual lymph mass > 1.5cm transverse diameter that has regressed > 75% or indeterminate BMB||- CRu by IWC with a completely negative PET but with an indeterminate BMB|
|PR||- ≥ 50% reduction in SPD of the six largest dominant nodes or nodal masses|
- no increase in size of spleen, liver, or other nodes
- no new sites of disease
|- CR, CRu, or PR by IWC with a positive PET at the site of a previously involved node/nodal mass|
- CR, CRu, PR, or SD by IWC with a positive PET outside the site of a previously involved node/nodal mass
- SD by IWC with a positive PET at the site of a previously involved node/nodal mass that regressed to < 1.5 cm if previously > 1.5 cm, or < 1 cm if previously 1.1-1.5 cm
|SD||- less than PR but not PD||- SD by IWC with a positive PET at the site of a previously involved node/nodal mass|
|PD||- applies only to patients with PR or nonresponders|
- ≥ 50% increase in the SPD from nadir of any previously identified abnormal node
- any new lesion
|- PD by IWC with a positive PET finding corresponding to the CT abnormality (new lesion, increasing size of previous lesion)|
- PD by IWC with a negative PET and a CT abnormality (new lesion, increasing size of previous lesion) of < 1.5 cm (< 1.0 cm in the lungs)
|RD||- applies only to patients with CR or Cru|
- ≥ 50% increase in size of previously involved sites or
- ≥ 50% increase in greatest diameter of any previously identified node > 1cm in short axis or
- ≥ 50% increase in the SPD of ≥ 2 nodes or
- any new lesion
|CR||Disappearance of all evidence of disease||(a) FDG-avid or PET positive prior to therapy; mass of any size permitted if PET negative|
(b) Variably FDG-avid or PET negative; regression to normal size on CT
|Not palpable, nodules disappeared||Inﬁltrate cleared on repeat biopsy; if indeterminate by morphology, immunohistochemistry|
|PR||Regression of measurable disease and no new sites||≥ 50% decrease in SPD of up to 6 largest dominant masses; no increase in size of other nodes|
(a) FDG-avid or PET positive prior to therapy; one or more PET positive at previously involved site
(b) Variably FDG-avid or PET negative; regression on CT
|≥ 50% decrease in SPD of nodules (for single nodule in greatest transverse diameter); no increase in size of liver or spleen||Irrelevant if positive prior to therapy; cell type should be speciﬁed|
|SD||Failure to attain CR/PR, or PD||(a) FDG-avid or PET positive prior to therapy; PET positive at prior sites of disease and no new sites on CT or PET|
(b) Variably FDG-avid or PET negative; no change in size of previous lesions on CT
|Relapsed Disease or PD||Any new lesion or increase of previously involved sites by ≥ 50% from nadir.||Appearance of a new lesion(s) > 1.5 cm in any axis, 50% increase in SPD of more than one node, or > 50% increase in longest diameter of a previously identified node >1 cm in short axis|
Lesions PET positive if FDG-avid lymphoma or PET positive prior to therapy.
|> 50% increase from nadir in the SPD of any previous lesions||New or recurrent involvement|
|Complete resolution of 18F-FDG uptake within tumor volume so that it was indistinguishable from surrounding normal tissue.||Complete resolution of 18F-FDG uptake within measurable target lesion so that it is less than mean liver activity and indistinguishable from surrounding background blood-pool levels.|
No new 18F-FDG-avid lesions in pattern typical of cancer.
Disappearance of all other lesions to background blood-pool levels.
|Percent reduction in SUVlbw should be recorded from measurable region and time (weeks) after treatment initiated|
(i.e., CMR 290, 4).
If anatomic progression by RECIST, must verify with follow-up.
|Reduction of minimum of 15% ± 25% in tumor 18F-FDG SUV after 1 cycle of chemotherapy, and >25% after >1 treatment cycle.|
Reduction in extent of tumor 18F-FDG uptake is not a requirement for PR.
|≥ 0.8 and ≥ 30% reduction of Peak* 18F-FDG SUVlbw in target measurable tumor.|
No new lesions.
SUVlbw measurement is obtained from the most active lesion also present at baseline (even if a different lesion than measured at baseline).
No increase > 30% in SUVlbw or size of target or nontarget lesions.
|Measurement is of the single most active lesion after treatment that was also present at baseline (e.g. may be a different lesion). Percent reduction in SUVlbw should be recorded and time in weeks after treatment initiated|
(i.e., PMR -40, 3).
If anatomic progression by RECIST, must verify with follow-up.
Reduction in extent of tumor 18F-FDG uptake is not requiremed.
|Increase in tumor 18F-FDG SUV <25% or decrease of < 15% and no visible increase in extent of 18F-FDG tumor uptake (20% in longest dimension).||No CMR, PMR, or PMD.||Peak SUVlbw in metabolic target lesion should be recorded, as well as time (weeks) from initation of most recent therapy, in weeks|
(i.e., SMD -15, 7).
|Increase in 18F-FDG tumor SUV of >25% within tumor region defined on baseline scan; visible increase in extent of 18F-FDG tumor uptake (20% in longest dimension) or appearance of new 18F-FDG uptake in metastatic lesions.||(1) >30% and >0.8 increase in 18F-FDG Peak* SUVlbw from baseline in pattern typical of tumor and not of infection/treatment effect.|
(2) Visible increase in extent of 18F-FDG tumor uptake (75% in TGA volume with no decline in SUVlbw
(3) New 18F-FDG-avid lesions that are typical of cancer and not related to treatment effect or infection.
|PD other than new visceral lesions should be confirmed on follow-up study within 1 month unless clearly associated with PD by RECIST 1.1.|
Should report percent change in Peak SUVlbw, time elased since treatment (weeks) and whether new lesions are present/absent and their number
(i.e., PMD, 135, 4, new: 5).
4. Clinical relevance of treatment response assessment
4.1. Head & neck cancer – Definitive/preoperative chemoradiation
18F-FDG PET has found a particularly significant role in treatment of head and neck cancers. It has long shown promise in its ability to prognosticate; in 37 patients from 1991-1994 with head and neck squamous cell carcinomas (HNSCC) receiving baseline 18F-FDG PET, SUVmax showed correlation with aggressive disease and potential prediction for survival .
Beyond prognostication, 18F-FDG PET is now routinely used to adapt treatment management, particularly in obviating surgical neck dissection in patients with complete response to initial radiation or chemoradiation therapy. Early studies have supported observation and omission of planned dissection after definitive radiotherapy for node-positive HNSCC with complete response on CT imaging, though at least selective nodal dissection was routinely practiced for residual neck masses [82,83]. With implementation of 18F-FDG PET, its negative predictive value has further supported omission of planned neck dissection, including in patients with residual neck mass/lymphadenopathy [84–88].
In an early study by Yao
In a further analysis, Yao
Such studies support timing of follow-up 18F-FDG PET to be 12 weeks post-treatment [84,85,88]. High negative predictive value (91%) has been shown at 16 weeks  post-treatment, though early time points (e.g. 4 weeks) have shown increased false positives . Metaanalyses support PET ≥ 12 weeks after completion of definitive therapy for moderately higher diagnostic accuracy. An added benefit of 18F-FDG PET at this early follow-up interval is the potential to spare neck dissection in patients who show early distant metastatic disease [88,89].
Despite lack of any randomized prospective studies, significant retrospective evidence has continued to show similar findings. Recent metaanalyses [90–92], discuss 26, 27, and 51 studies including up to 2335 patients , overall supporting the high negative predictive value (approximately 95%) of follow-up PET and its value in omitting planned neck dissection. Further, despite the increased costs of PET imaging, PET-guided management in patients with complete response at the primary site has shown to be the more cost effective than CT-guided management or planned neck dissection .
4.2. Rectal cancer – Preoperative chemoradiation
Similar to HNSCC, first line treatment for locally-advanced rectal cancer includes upfront chemoradiation. In this setting, however, subsequent planned surgery remains standard of care. This multimodality neoadjuvant approach has shown to decrease local recurrence and improve overall survival [94,95]. Furthermore, neoadjuvant treatment has shown to increase sphincter-preserving surgery, conferring decreased surgical morbidity and improved quality of life [96–98].
Deferring subsequent surgical intervention in this disease site has similarly been investigated. In a cohort of 71 patients with distal rectal carcinoma considered resectable prior to concurrent chemoradiation with subsequent complete clinical response treated subsequently with observation alone (no planned surgery), five-year overall and disease-free survivals were 100% and 92%, respectively.
Improving restaging methods after neoadjuvant chemotherapy provides clinicians with increased information to guide management. Radiographic imaging modalities, however, are less sensitive to assessment of pathologic response, which is better characterized by metabolic imaging with 18F-FDG PET [54,99,100].
A number of studies have attempted correlation of 18F-FDG PET with tumor downstaging and response to neoadjuvant chemoradiation [100–105]. In a study by Capirci
Notably, surgery is routinely planned approximately 6 weeks after neoadjuvant treatment, as surgery at 6-weeks was shown to have more tumor downstaging than at 2 weeks . However, further tumor response and increased survival has been noted with intervals > 7 weeks . A recent similar study by Perez
In rectal cancer, 18F-FDG PET restaging does show promise in potentially affecting treatment management; prospective studies investigating its role in this setting are awaited.
18FDG-PET finds various roles in management of lymphoma. For staging in Hodgkin lymphoma (HL) and non-Hodgkin lymphoma (NHL), PET with CT (PET-CT) has been shown to improve sensitivity and specificity in evaluation of nodal and extranodal sites in comparison to contrast-enhanced CT without PET [108,109]. It has further shown to be 92% sensitive for bone marrow involvement in HL . Beyond staging, PET has been used for post-chemotherapy restaging, assessing response during chemotherapy at initial diagnosis, and also during salvage treatment. In current NCCN guidelines for both HL & NHL, PET-CT has variably been incorporated into staging, restaging during chemotherapy, and restaging after chemotherapy; routine PET-CT in the surveillance setting, however, is recommended against secondary to false-positive risk [111,112].
The role for PET in lymphoma is clearest in the setting of restaging, either during or subsequent to treatment. PET has a very high negative predictive value (88-100%, see Table 6) . Further, after treatment, PET is superior to CT for distinguishing residual mass with versus without residual viable disease (e.g. post-treatment fibrosis) . Spaepen
More research has investigated interim (during chemotherapy) 18FDG-PET for assessment of treatment response and prognostication (see Table 7). Cerci
|Hutchings ||2005||85||40||2-3||74% PET-|
|97% 2y PFS|
46% 2y PFS
|Hutchings ||2006||77||23||2||79% PET-|
|96% 2y PFS|
0% 2y PFS
|Kostakoglu ||2006||23||21||1||74% PET-|
|100% 2y PFS|
13% 2y PFS
|Zinzani ||2006||40||18||2||80% PET-|
|Gallamini ||2007||260||26||2||81% PET-|
|95% 2y PFS|
13% 2y PFS
|Markova ||2009||50||25||4||72% PET-|
|Cerci ||2010||104||36||2||71% PET-|
|91% 3y EFS|
53% 3y EFS
|Jerusalem ||2000||28||18||3||82% PET-|
|62% 2y PFS|
0% 2y PFS
|Spaepen ||2002||70||36||2-3||53% PET -|
47% PET +
|Haioun ||2005||90||24||2||60% PET -|
40% PET +
|82% 2y EFS|
43% 2y EFS
|Mikhaell ||2005||121||29||2-3||41% PET -|
43% PET +
|88% 5y PFS|
16% 5y PFS
59% 5y PFS
|Ng ||2007||45||31||1-5||69% PET -|
31% PET +
|Han ||2009||40||24||2-4||68% PET -|
32% PET +
|Pregno ||2009||88||26||2-4||72% PET -|
28% PET +
|85% 2y PFS|
72% 2y PFS
|Safar ||2009||112||38||2||63% PET -|
37% PET +
|84% 3y PFS|
47% 3y PFS
|Cashen ||2011||50||15||2-3||52% PET -|
48% PET +
|85% 2y PFS|
63% 2y EFS
|Zinzani ||2011||91||50||variable||62% PET-|
|75% 4y EFS|
18% 4y EFS
In the setting of relapsing/refractory Hodgkin lymphoma, interim PET after 2 cycles of salvage high-dose chemotherapy has been assessed. Limited retrospective data from Castagna
In the German Hodgkin Study Group HD15 trial (2012) [119,140] with over 2,000 patients with advanced-stage Hodgkin lymphoma, 3 BEACOPP chemotherapy regimens were compared in a non-inferiority randomized trial. Radiotherapy was implemented with a “PET-guided” adaptive approach based on post-chemotherapy response regardless of treatment arm. If a PET-positive persistent mass 2.5cm or larger was present after completion of chemotherapy (median 21 days), 30Gy local radiation therapy was administered for consolidation. Negative predictive value for post-chemotherapy PET was 94% at 12 months follow-up. In the 3 arms, five-year freedom from failure ranged from 84%-89%, and five-year survival ranged from 92-95%. Consolidative radiotherapy was not randomized and was administered to 11% of patients (compared to 71% in HD9 ). With such excellent outcomes with this PET-guided radiotherapy approach, the authors indicate this approach as their current standard of care. Longer follow-up and prospective clinical trials assessing need for consolidative radiotherapy are still awaited.
4.4. Esophageal cancer – Definitive/preoperative
The role of multimodality therapy for esophageal and gastroesophageal cancer has historically not been well defined. Resection has been considered standard treatment for patients with resectable/localized disease without strong evidence supporting neoadjuvant therapy, despite significant risk for local and distant recurrences yielding poor 5-year survival rates ranging from 15-39%. Neoadjuvant treatment is increasingly becoming adopted as standard of care for locally-advanced disease, with use continuing to increase [143,144]. Multiple prospective trials did not report survival benefit with neoadjuvant chemoradiotherapy [145–147], and randomized studies supporting neoadjuvant treatment are scarce. Walsh
In patients receiving neoadjuvant chemoradiation, a portion—29% in the Dutch CROSS study—are found to have pathologic complete response on subsequent surgery. In a single-institution review, pathologic complete response from neoadjuvant treatment was associated with higher 5-year and overall survival (48% vs. 18% and 50 months vs. 28 months, respectively) in comparison to patients without complete response . With treatment response bearing significant prognostic potential, assessment of response to neoadjuvant treatment for esophageal cancer has been an area of increasing research [150–163].
In an early study by Weber
In a follow-up study , patients had three 18FDG-PET scans: one pretreatment, one during treatment (2 weeks after starting), then 3-4 weeks preoperatively (but after neoadjuvant treatment. Responders had more decrease at 2 weeks (44% vs. 21%, p<0.01) and preoperatively (70% vs. 51%, p=0.01). During-treatment PET had higher power than the preoperative PET treatment to predict response (area under curve (AUC) of receiver operator characteristic (ROC) 0.78 vs. 0.88), though difference was not statistically significant (p=0.40). Best cutoff for response in this cohort was 30% reduction from baseline (93% sensitive, 88% accurate), who all proceed to have R0 resection. Responders by this PET criteria had higher survival (median 38 vs. 18 months; 2-year rates 79% vs. 38%, p<0.01).
Analysis of gastroesophageal junction tumors again showed improved prognostic potential with PET using percent reduction of SUVmax 2 weeks after treatment start (p=0.03) versus after completion of neoadjuvant treatment (p=0.09) . Though percent reduction is routinely used to assess response, thresholds of decrease of SUVmax (e.g. decrease of ≥10) from before to after neoadjuvant treatment have shown to predict significant histopathologic response .
More recent studies have showed other metrics as better predictors of response. In a comparison of SUVmax, MTV based on fixed threshold of 2.5 SUV, and SUVmean (of MTV), and TGA, MTV and TGA were both 91% sensitive in predicting histopathologic response when also using CT, but MTV increase specificity from 90% to 93%. Most predictive was TGA (AUC=0.95) followed by MTV (AUC=0.92), SUVmax (AUC=0.84), and SUVmean (AUC=0.82) . Further, metabolic response criteria (e.g. PERCIST) have shown better assessed response in comparison to non-metabolic methods (e.g. RECIST and WHO) [159,163].
With various studies showing prognostic potential of 18FDG-PET early during treatment, there is question as to the utility of PET to potentially facilitate treatment modification . Kwee (2010)  performed a metaanalysis of 20 PET-response studies including 849 patients; it however showed wide ranges of sensitivity and specificity with overall AUC of 0.78. Based on the pooled data, PET was not recommended for routine clinical use to guide neoadjuvant treatment. Furthermore, in a retrospective single-institution review , patients treated with neoadjuvant chemotherapy followed by surgery had similar freedom from local failure (p=0.92) and overall survival (p=0.15) in comparison to patients receiving definitive chemoradiation who attained metabolic CR (SUV<3). Furthermore, in this retrospective study, though not statistically significant, rate of death in the definitive chemoradiation group was higher than in the surgical group despite worse baseline characteristics.
Similar to head and neck cancer, prospective studies are awaited to formally assess necessity of surgical management after complete metabolic response to neoadjuvant chemoradiation therapy in operable/resectable patients.
4.5. Non-Small Cell Lung Cancer (NSCLC)
18FDG-PET is currently recommended by NCCN guidelines for routine staging of stage I-III NSCLC . Radiotherapy planning with PET fusion has further been recommended for biologically-targeted radiotherapy in which 3D-PET fusion is implemented for tumor delineation, with PET performed with minimal delay between PET and start of treatment, given propensity for rapid disease progression [64–66,165]. Metabolic (PET) response to treatment has been shown to pre-date radiographic (CT) response. Despite increasing data showing utility of PET for assessing treatment response in NSCLC and predicting outcomes including survival, guidelines currently do not recommend PET in this setting [44,45,67,76,166–177].
In an early study of 15 patients receiving chemotherapy for IIIB-IV NSCLC, patients received weekly PET starting at initiation of chemotherapy until completion of 2 cycles (6 weeks later) . Reduction of SUVmax by 50% week 1 to week 3 was predictive of survival of > 6 months, thus facilitating prediction of response to treatment. Those with less reduction died within 6 months. In patients without early response, management may thus be altered to forego futile chemotherapy. In an early study  of 15 stage I-III patients receiving radiotherapy, patients received 3 PETs: one pre-treatment, one during treatment after approximately 45 Gy, and one 3 months post-treatment. Response during treatment was shown to correlate with overall response after treatment (p=0.03), and SUV during treatment correlated with SUV 3 months after (p<0.001). A number of studies with prospective PET data with cutoffs are listed in Table 8.
|Vansteenkiste ||1985||15||IIIA||50% decrease||OS||0.03|
|Weber ||2003||57||IIIB-IV||20% decrease||OS||<0.01|
|Hellwig ||2004||47||IIB-III||SUV < 4||OS||<0.01|
|Eschmann ||2007||70||III||CMR or 80% decrease||OS||<0.01|
|de Geus-Oei ||2007||51||IB-IV||35% decrease||OS||0.02|
|Nahmias ||2007||16||IIIB-IV||50% decrease from week 1 to week 3||OS||<0.01|
|At 12 months|
|Mangona ||2012||129||IA-IB||SUV ≥ 3.9|
SUV ≥ 6.0
|Mangona ||2012||16||IIB-IIIB||30% decrease|
decrease ≥ 4
Stereotactic body radiotherapy (SBRT), employing modern techniques including 4-D treatment planning and image-guided radiotherapy (IGRT) has been shown to be an effective, cost-efficient, treatment option for definitive management of early-stage NSCLC as well as lung metastases from other organs with excellent tumor control rates; in comparison to medically-operable patients who are treated with resection, retrospective data of primarily medically-inoperable patients with poor pulmonary function suggests excellent tumor control with SBRT with rates similar to that of sublobar resection and minimal toxicity [179–191].
In a large single-institution analysis  of 129 consecutive NSCLC tumors treated with SBRT, 58% enrolled on a prospective phase II protocol, patients had baseline and serial follow-up PET imaging. Sixteen patients additionally had weekly on-treatment 4D-PET-CT. Median follow-up was 19 months and median time until local failure (LF) of 15 months. A total of 475 PETs were obtained. Change in SUV from pre-treatment to follow-up are seen in Figure 1 and stratified by status of LF vs. no-LF based on last follow-up. Though baseline SUVmax was higher in the LF group (12.4 vs. 6.5, p=0.0001), difference was not significant at 1.5 and 6 months, as both groups responded. SUV at 12 months, however, was significantly higher for the LF vs. no-LF group (6.8 vs. 2.5, p=0.02). Cutoffs predictive of LF were 12-month SUV ≥ 3.9 (100 sensitive), 12-month SUV ≥ 6 (100 specific), and 12-month SUV ≥ 40% of baseline (see Table 8). Analysis of SUVmax velocity showed trend for higher velocity at 12 months (+0.18 SUV/month vs. -0.03 SUV/month, p=0.058). On multivariate logistic regression, 12-month SUV was most predictive of LF (p=0.057).
In a cohort of 16 patients with locally-advanced NSCLC enrolled on a phase II protocol, patients had PET at baseline, weekly during treatment, and at follow-up [70,192] (see Figure 2). Patients received hyperfractionated radiation therapy 1.5 Gy BID with concurrent chemotherapy either as definitive treatment (n=12) or as neoadjuvant treatment (n=4) delivering RT with daily online cone-beam CT for image guidance and intensity modulated radiotherapy (IMRT) to minimize potential normal tissue toxicity [190,193,194]. After potential follow-up of 20 months (range 12-28), 7 had locoregional recurrence (LRR), and 8 died (5 of disease). Interestingly, there was trend for higher SUVmax at baseline in those without LRR (the no-LRR group) than in those with LRR (19.0 vs. 11.9, p=0.08), an inverse relationship than expected. The rate of SUV decrease in the LRR group during RT was 1.6 per week, significantly faster than the no-LRR group (0.23 per week, p=0.02) such that SUV values were similar for both groups by the 4th on-treatment PET (p=0.95) (see Table 9). A during-RT decrease of less than 4 from baseline was predictive of LRR (p<0.01), and a during-RT decrease less 30% from baseline was predictive of death from disease (p<0.01). Velocity of retention index from PET1 to PET-FU predicted overall survival (+1.6%/week in those who died vs. -1.7%/week in those alive, p=0.03).
PET shows prognostic potential in this disease site from prior to treatment to early in treatment, to later in follow-up. It further holds potential for adjusting management (e.g. discontinuing ineffective chemotherapy, potentially modifying radiation therapy during treatment, and predicting delayed local failure for potential earlier biopsy/intervention). We await further prospective PET data and clinical trials to best define the role of PET in assessment of treatment in NSCLC.
5. Future directions
As PET is used for staging and radiotherapy prior to treatment for a number of organ sites, PET further has potential for restaging and replanning radiotherapy during the course of therapy. Beyond mid-treatment prognostication, this facilitates potential treatment modification. For radiotherapy re-planning, potential changes are include modification of target volumes based on anatomic changes from treatment, modification of boost volumes, and potentially adjustment of prescription dose based on response (e.g. higher dose for poor responders vs. less dose for good responders). Such investigations are currently ongoing in clinical protocols.
In treatment of locally-advanced head and neck squamous cell carcinomas, our institution has initiated a prospective, non-randomized trial evaluating the utility of such an adaptive approach focusing on target volume adaptation. Patients receiving 70 Gy IMRT in 35 daily fractions (7 week duration) with concurrent cisplatin or cetuximab are eligible. 18FDG-PET-CT is utilized for treatment planning. Repeat PET-CTs and diagnostic CTs are obtained after fractions 10 and 22 for the purpose of treatment adaptation. Three different treatment plans will be created, one for fractions 1-12 (based on pre-treatment PET-CT), one for fractions 13-24 (based on PET-CT after fraction 10), and one for fractions 25-35 (based on PET-CT after fraction 22). Such an adaptive approach may help decrease dose delivered to normal tissue as tumors decrease in size during treatment, potentially decreasing toxicity. On this protocol, patients also obtain weekly PET-CTs for assessment of treatment response, though prescription dose is not modified in this study.
For non-small cell lung cancer, investigators have further used on-treatment PET to facilitate PET-adaptive replanning, with PET-adaptive dose escalation incorporated into a currently-enrolling Radition Therapy Oncology Group (RTOG) Protocol, RTOG 1106 [195,196]. All patients on this protocol will have 18FDG-PET; however, a subset are planned to also have 18F-MISO-PET at staging.
As such radiotracers beyond 18F-FDG show particular promise, further results of clinical trials implementing these are awaited.
Over the past 20 years, the body of data assessing treatment response with PET has grown significant. Assessing treatment response with PET can yield highly prognostic information. Such information, however, may have no end-effect on management. As clinicians, many of our PET-based decisions are based on retrospective and prospective data without comparison of management options based on PET results. Such results are significantly hypothesis-generating. The high negative predictive value of PET in various organ sites may increase comfort of clinicians when considering omitting potentially unnecessary interventions (e.g. neck dissection after complete metabolic response of locally-advanced head and neck cancer to chemoradiation, esophagectomy after complete metabolic response to chemoradiotion, or consolidative radiotherapy after complete metabolic response in Hodgkin lymphoma). High-level evidence to justify such treatment-adapting decisions based on PET are currently lacking, thus we caution application of such data as justification for modifying standard of care. We strongly encourage PET-adaptive management under the guise of clinical trials at this time, as the role of PET in oncology continues to best be defined.
Special thanks to Dr. Katie Traylor, Rob Ceruti, and Matthew Johnson of Nuclear Medicine for their assistance with PET figures presented in this chapter.