The number of target lesions to be assessed was reduced from five per organ to two per organ and from a maximum of 10 target lesions total to a maximum of five total [2] (Figs. 6A, 6B and 7A, 7B). The change is based on the analysis of a large prospective database from 16 clinical trials that showed that assessment of five lesions per patient did not influence the overall response rate and only minimally affected progression-free survival [3].

In RECIST 1.0, there was no clear guideline for lymph node measurement. In RECIST 1.1, detailed instructions about how to measure and assess lymph nodes are provided [2, 5]. Lymph nodes with a short axis of ≥ 15 mm are considered measurable and assessable as target lesions, and the short-axis measurement should be included in the sum of target lesion measurements in the calculation of tumor response as opposed to the longest axis used for measurements of other target lesions. Lymph nodes with a short axis of < 10 mm are defined as “nonpathologic” (Fig. 8A). All other pathologic nodes—that is, those with a short axis of ≥ 10 mm but < 15 mm—should be considered nontarget lesions [2] (Fig. 8B).

Given the changes made in the assessment of pathologic lymph nodes, CR by RECIST 1.1 requires, first, the disappearance of all target lesions; and, second, a reduction in the short-axis measurement of all pathologic lymph nodes (whether target or nontarget) to < 10 mm [2]. It is likely that the changes in the lymph node assessment mainly influence the evaluation of patients with epithelial cancers, which tend to metastasize to the lymph nodes.

PD for target lesions according to RECIST 1.1 requires a 5-mm absolute increase of the sum of the longest diameters of the target lesions in addition to a 20% increase in the sum of the target lesions [2] (Fig. 9). The new criterion of a 5-mm absolute change in size is particularly important in the follow-up assessment of patients with small-volume disease after response to therapy because a minimal increase in size due to measurement variability could meet the criterion of 20% increase by RECIST 1.0 without a true increase in tumor burden in these patients (Fig. 10A, 10B, 10C, 10D).

RECIST 1.1 further clarifies when one can assign “unequivocal progression” of nontarget lesions in the response assessment [2]. When measurable disease or a target lesion is present, the overall level of substantial worsening in nontarget disease, which leads to an increase of overall tumor burden even with SD or PR in target disease, is required to assign progression [2]. RECIST 1.1 emphasizes that a modest increase in the size of one or more nontarget lesions is usually not sufficient and that progression solely based on change in nontarget disease in patients with SD or PR of target disease is extremely rare [2].

If no measurable disease is present, which may be the case in some phase III trials that do not require measurable disease as a criterion for study entry, the same general concepts apply as in cases with measurable disease [2]. An increase of tumor burden that would be required to declare PD for measurable disease should be present. Examples include an increase in pleural effusion from trace to large or an increase in lymphangitic disease from localized to widespread [2].

One of the major changes in RECIST 1.1 is the inclusion of FDG PET in the detection of new lesions that define progression [2]. RECIST 1.1 provides a detailed guideline for using FDG PET to detect new lesions, as summarized in Figure 11. The inclusion of FDG PET, which evaluates glucose metabolism of tumor, adds a new functional aspect of response assessment to RECIST, which has been solely based on morphologic assessment using size measurements (Fig. 12A, 12B).

In patients with GIST naive to any molecular targeted therapy and initially treated with the tyrosine kinase inhibitor imatinib, CT attenuation has been used as an important tumor response parameter. After treatment with imatinib, the overall tumor CT attenuation decreases dramatically with development of myxoid degeneration, hemorrhage, or necrosis [7, 8] (Fig. 13A, 13B). This pattern of response is important to recognize particularly in the context of hepatic metastases because apparent “new lesions” may appear in response to therapy. These lesions are initially isodense to the hepatic parenchyma but become hypodense (and therefore visible) in response to therapy and should not be confused with new lesions and misinterpreted for PD. If there is any ambiguity regarding the CT interpretation, FDG PET can help resolve this and confirm that there is no metabolic activity within these masses. Choi response criteria, defined as a 10% decrease in the unidimensional tumor size or a 15% decrease in CT attenuation, have been shown to correlate well with response by FDG PET and to be more predictive of time to progression than response by RECIST 1.0 [8]. Similar criteria using CT attenuation have been shown to be useful for response assessment in soft-tissue sarcomas other than GIST and in carcinomas [9, 10]. It is important to note that, in patients with GIST after the initial response to imatinib, recurrent disease may show a “nodule-within-a-mass” pattern that cannot be detected as PD using conventional size measurements alone [11].

Targeted anticancer therapy using antiangiogenesis agents or tyrosine kinase inhibitors is known to cause a paradoxical increase of tumor size despite response because of hemorrhage or necrosis. This phenomenon is often seen in liver tumors, including hepatocellular carcinoma and liver metastasis from GIST or melanoma [11, 12]. A paradoxical increase in size in the setting of hemorrhage or necrosis in responding tumors should not be mistaken for PD, and MRI can be performed to confirm the presence of these intratumoral changes (Fig. 14A, 14B, 14C, 14D). FDG PET can also help confirm metabolic response in these tumor masses despite the apparent increase in size. Radiologists must be aware of this phenomenon to avoid misinterpretation and to prompt appropriate further evaluation by MRI or PET/CT.

Cavitation of lung lesions is commonly observed, especially in non–small cell lung cancer treated with antiangiogenic agents such as vascular endothelial growth factor receptor inhibitors [13, 14]. Cavitation of lung lesions provides a challenge to radiologists who try to obtain the appropriate measurement that best represents tumor burden. Rather than simply measuring the longest diameter of a lesion including the area of cavitation, an alternative measurement that excludes the area of cavitation has been proposed [14] (Fig. 15). Response assessment might be improved by incorporating cavitation into tumor measurement; however, further prospective study is needed to validate this hypothesis.